mesos.proto

// Licensed to the Apache Software Foundation (ASF) under one
// or more contributor license agreements.  See the NOTICE file
// distributed with this work for additional information
// regarding copyright ownership.  The ASF licenses this file
// to you under the Apache License, Version 2.0 (the
// "License"); you may not use this file except in compliance
// with the License.  You may obtain a copy of the License at
//
//     http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.

package mesos;

option java_package = "org.apache.mesos";
option java_outer_classname = "Protos";


/**
 * Status is used to indicate the state of the scheduler and executor
 * driver after function calls.
 */
enum Status {
  DRIVER_NOT_STARTED = 1;
  DRIVER_RUNNING = 2;
  DRIVER_ABORTED = 3;
  DRIVER_STOPPED = 4;
}


/**
 * A unique ID assigned to a framework. A framework can reuse this ID
 * in order to do failover (see MesosSchedulerDriver).
 */
message FrameworkID {
  required string value = 1;
}


/**
 * A unique ID assigned to an offer.
 */
message OfferID {
  required string value = 1;
}


/**
 * A unique ID assigned to a slave. Currently, a slave gets a new ID
 * whenever it (re)registers with Mesos. Framework writers shouldn't
 * assume any binding between a slave ID and and a hostname.
 */
message SlaveID {
  required string value = 1;
}


/**
 * A framework-generated ID to distinguish a task. The ID must remain
 * unique while the task is active. A framework can reuse an ID _only_
 * if the previous task with the same ID has reached a terminal state
 * (e.g., TASK_FINISHED, TASK_LOST, TASK_KILLED, etc.). However,
 * reusing task IDs is strongly discouraged (MESOS-2198).
 */
message TaskID {
  required string value = 1;
}


/**
 * A framework-generated ID to distinguish an executor. Only one
 * executor with the same ID can be active on the same slave at a
 * time. However, reusing executor IDs is discouraged.
 */
message ExecutorID {
  required string value = 1;
}


/**
 * ID used to uniquely identify a container. If the `parent` is not
 * specified, the ID is a UUID generated by the agent to uniquely
 * identify the container of an executor run. If the `parent` field is
 * specified, it represents a nested container.
 */
message ContainerID {
  required string value = 1;
  optional ContainerID parent = 2;
}


/**
 * Represents time since the epoch, in nanoseconds.
 */
message TimeInfo {
  required int64 nanoseconds = 1;

  // TODO(josephw): Add time zone information, if necessary.
}


/**
 * Represents duration in nanoseconds.
 */
message DurationInfo {
  required int64 nanoseconds = 1;
}


/**
 * A network address.
 *
 * TODO(bmahler): Use this more widely.
 */
message Address {
  // May contain a hostname, IP address, or both.
  optional string hostname = 1;
  optional string ip = 2;

  required int32 port = 3;
}


/**
 * Represents a URL.
 */
message URL {
  required string scheme = 1;
  required Address address = 2;
  optional string path = 3;
  repeated Parameter query = 4;
  optional string fragment = 5;
}


/**
 * Represents an interval, from a given start time over a given duration.
 * This interval pertains to an unavailability event, such as maintenance,
 * and is not a generic interval.
 */
message Unavailability {
  required TimeInfo start = 1;

  // When added to `start`, this represents the end of the interval.
  // If unspecified, the duration is assumed to be infinite.
  optional DurationInfo duration = 2;

  // TODO(josephw): Add additional fields for expressing the purpose and
  // urgency of the unavailability event.
}


/**
 * Represents a single machine, which may hold one or more slaves.
 *
 * NOTE: In order to match a slave to a machine, both the `hostname` and
 * `ip` must match the values advertised by the slave to the master.
 * Hostname is not case-sensitive.
 */
message MachineID {
  optional string hostname = 1;
  optional string ip = 2;
}


/**
 * Holds information about a single machine, its `mode`, and any other
 * relevant information which may affect the behavior of the machine.
 */
message MachineInfo {
  // Describes the several states that a machine can be in.  A `Mode`
  // applies to a machine and to all associated slaves on the machine.
  enum Mode {
    // In this mode, a machine is behaving normally;
    // offering resources, executing tasks, etc.
    UP = 1;

    // In this mode, all slaves on the machine are expected to cooperate with
    // frameworks to drain resources.  In general, draining is done ahead of
    // a pending `unavailability`.  The resources should be drained so as to
    // maximize utilization prior to the maintenance but without knowingly
    // violating the frameworks' requirements.
    DRAINING = 2;

    // In this mode, a machine is not running any tasks and will not offer
    // any of its resources.  Slaves on the machine will not be allowed to
    // register with the master.
    DOWN = 3;
  }

  required MachineID id = 1;
  optional Mode mode = 2;

  // Signifies that the machine may be unavailable during the given interval.
  // See comments in `Unavailability` and for the `unavailability` fields
  // in `Offer` and `InverseOffer` for more information.
  optional Unavailability unavailability = 3;
}


/**
 * Describes a framework.
 */
message FrameworkInfo {
  // Used to determine the Unix user that an executor or task should be
  // launched as.
  //
  // When using the MesosSchedulerDriver, if the field is set to an
  // empty string, it will automagically set it to the current user.
  //
  // When using the HTTP Scheduler API, the user has to be set
  // explicitly.
  required string user = 1;

  // Name of the framework that shows up in the Mesos Web UI.
  required string name = 2;

  // Note that 'id' is only available after a framework has
  // registered, however, it is included here in order to facilitate
  // scheduler failover (i.e., if it is set then the
  // MesosSchedulerDriver expects the scheduler is performing
  // failover).
  optional FrameworkID id = 3;

  // The amount of time (in seconds) that the master will wait for the
  // scheduler to failover before it tears down the framework by
  // killing all its tasks/executors. This should be non-zero if a
  // framework expects to reconnect after a failure and not lose its
  // tasks/executors.
  //
  // NOTE: To avoid accidental destruction of tasks, production
  // frameworks typically set this to a large value (e.g., 1 week).
  optional double failover_timeout = 4 [default = 0.0];

  // If set, framework pid, executor pids and status updates are
  // checkpointed to disk by the slaves. Checkpointing allows a
  // restarted slave to reconnect with old executors and recover
  // status updates, at the cost of disk I/O.
  optional bool checkpoint = 5 [default = false];

  // Used to group frameworks for allocation decisions, depending on
  // the allocation policy being used.
  optional string role = 6 [default = "*"];

  // Used to indicate the current host from which the scheduler is
  // registered in the Mesos Web UI. If set to an empty string Mesos
  // will automagically set it to the current hostname if one is
  // available.
  optional string hostname = 7;

  // This field should match the credential's principal the framework
  // uses for authentication. This field is used for framework API
  // rate limiting and dynamic reservations. It should be set even
  // if authentication is not enabled if these features are desired.
  optional string principal = 8;

  // This field allows a framework to advertise its web UI, so that
  // the Mesos web UI can link to it. It is expected to be a full URL,
  // for example http://my-scheduler.example.com:8080/.
  optional string webui_url = 9;

  message Capability {
    enum Type {
      // This must be the first enum value in this list, to
      // ensure that if 'type' is not set, the default value
      // is UNKNOWN. This enables enum values to be added
      // in a backwards-compatible way. See: MESOS-4997.
      UNKNOWN = 0;

      // Receive offers with revocable resources. See 'Resource'
      // message for details.
      REVOCABLE_RESOURCES = 1;

      // Receive the TASK_KILLING TaskState when a task is being
      // killed by an executor. The executor will examine this
      // capability to determine whether it can send TASK_KILLING.
      TASK_KILLING_STATE = 2;

      // Indicates whether the framework is aware of GPU resources.
      // Frameworks that are aware of GPU resources are expected to
      // avoid placing non-GPU workloads on GPU agents, in order
      // to avoid occupying a GPU agent and preventing GPU workloads
      // from running! Currently, if a framework is unaware of GPU
      // resources, it will not be offered *any* of the resources on
      // an agent with GPUs. This restriction is in place because we
      // do not have a revocation mechanism that ensures GPU workloads
      // can evict GPU agent occupants if necessary.
      //
      // TODO(bmahler): As we add revocation we can relax the
      // restriction here. See MESOS-5634 for more information.
      GPU_RESOURCES = 3;

      // Receive offers with resources that are shared.
      // TODO(anindya_sinha): This is currently a no-op.
      SHARED_RESOURCES = 4;

      // Indicates that the framework is prepared to handle the
      // following TaskStates: TASK_UNREACHABLE, TASK_DROPPED,
      // TASK_GONE, TASK_GONE_BY_OPERATOR, and TASK_UNKNOWN.
      //
      // With this capability, frameworks can define how they would
      // like to handle partitioned tasks. Frameworks will receive
      // TASK_UNREACHABLE for tasks on partitioned agents; if/when the
      // partitioned agent reregisters, the task will not be killed.
      // Frameworks that enable this capability will never receive
      // TASK_LOST; they will receive one of the most specific task
      // statuses listed above instead.
      //
      // Without this capability, frameworks will receive TASK_LOST
      // for tasks on partitioned agents; such tasks will be killed by
      // Mesos when the agent reregisters (unless the master has
      // failed over).
      PARTITION_AWARE = 5;
    }

    // Enum fields should be optional, see: MESOS-4997.
    optional Type type = 1;
  }

  // This field allows a framework to advertise its set of
  // capabilities (e.g., ability to receive offers for revocable
  // resources).
  repeated Capability capabilities = 10;

  // Labels are free-form key value pairs supplied by the framework
  // scheduler (e.g., to describe additional functionality offered by
  // the framework). These labels are not interpreted by Mesos itself.
  // Labels should not contain duplicate key-value pairs.
  optional Labels labels = 11;
}


/**
 * Describes a health check for a task or executor (or any arbitrary
 * process/command). A "strategy" is picked by specifying one of the
 * optional fields. Specifying more than one strategy is an error.
 */
message HealthCheck {
  enum Type {
    UNKNOWN = 0;
    COMMAND = 1;
    HTTP = 2;
    TCP = 3;
  }

  // Describes an HTTP health check. Sends a GET request to
  // scheme://<host>:port/path. Return codes between 200 and 399 are treated
  // as success, failure otherwise. Note that <host> is not configurable
  // and is resolved automatically, in most cases to localhost.
  message HTTPCheckInfo {
    // Currently "http" and "https" are supported.
    optional string scheme = 1;

    // Port to send the HTTP request.
    required uint32 port = 2;

    // HTTP request path.
    optional string path = 3;

    // TODO(alexr): Add support for HTTP method. While adding POST
    // and PUT is simple, supporting payload is more involved.

    // TODO(alexr): Add support for custom HTTP headers.

    // TODO(alexr): Add support for success and possibly failure
    // statuses.

    // TODO(haosdent): Consider adding a flag to enable task's certificate
    // validation for HTTPS health checks, see MESOS-5997.

    // TODO(benh): Include an 'optional bytes data' field for checking
    // for specific data in the response.
  }

  // Describes a TCP health check, i.e. based on establishing
  // a TCP connection to the specified port.
  message TCPCheckInfo {
    // Port expected to be open.
    required uint32 port = 1;
  }

  // TODO(benh): Consider adding a URL health check strategy which
  // allows doing something similar to the HTTP strategy but
  // encapsulates all the details in a single string field.

  // Amount of time to wait until starting the health checks.
  optional double delay_seconds = 2 [default = 15.0];

  // Interval between health checks.
  optional double interval_seconds = 3 [default = 10.0];

  // Amount of time to wait for the health check to complete.
  optional double timeout_seconds = 4 [default = 20.0];

  // Number of consecutive failures until signaling kill task.
  optional uint32 consecutive_failures = 5 [default = 3];

  // Amount of time to allow failed health checks since launch.
  optional double grace_period_seconds = 6 [default = 10.0];

  // TODO(alexr): Add an optional `KillPolicy` that should be used
  // if the task is killed because of a health check failure.

  // The type of health check.
  optional Type type = 8;

  // Command health check.
  optional CommandInfo command = 7;

  // HTTP health check.
  optional HTTPCheckInfo http = 1;

  // TCP health check.
  optional TCPCheckInfo tcp = 9;
}


/**
 * Describes a kill policy for a task. Currently does not express
 * different policies (e.g. hitting HTTP endpoints), only controls
 * how long to wait between graceful and forcible task kill:
 *
 *     graceful kill --------------> forcible kill
 *                    grace_period
 *
 * Kill policies are best-effort, because machine failures / forcible
 * terminations may occur.
 *
 * NOTE: For executor-less command-based tasks, the kill is performed
 * via sending a signal to the task process: SIGTERM for the graceful
 * kill and SIGKILL for the forcible kill. For the docker executor-less
 * tasks the grace period is passed to 'docker stop --time'.
 */
message KillPolicy {
  // The grace period specifies how long to wait before forcibly
  // killing the task. It is recommended to attempt to gracefully
  // kill the task (and send TASK_KILLING) to indicate that the
  // graceful kill is in progress. Once the grace period elapses,
  // if the task has not terminated, a forcible kill should occur.
  // The task should not assume that it will always be allotted
  // the full grace period. For example, the executor may be
  // shutdown more quickly by the agent, or failures / forcible
  // terminations may occur.
  optional DurationInfo grace_period = 1;
}


/**
 * Describes a command, executed via: '/bin/sh -c value'. Any URIs specified
 * are fetched before executing the command.  If the executable field for an
 * uri is set, executable file permission is set on the downloaded file.
 * Otherwise, if the downloaded file has a recognized archive extension
 * (currently [compressed] tar and zip) it is extracted into the executor's
 * working directory. This extraction can be disabled by setting `extract` to
 * false. In addition, any environment variables are set before executing
 * the command (so they can be used to "parameterize" your command).
 */
message CommandInfo {
  message URI {
    required string value = 1;
    optional bool executable = 2;

    // In case the fetched file is recognized as an archive, extract
    // its contents into the sandbox. Note that a cached archive is
    // not copied from the cache to the sandbox in case extraction
    // originates from an archive in the cache.
    optional bool extract = 3 [default = true];

    // If this field is "true", the fetcher cache will be used. If not,
    // fetching bypasses the cache and downloads directly into the
    // sandbox directory, no matter whether a suitable cache file is
    // available or not. The former directs the fetcher to download to
    // the file cache, then copy from there to the sandbox. Subsequent
    // fetch attempts with the same URI will omit downloading and copy
    // from the cache as long as the file is resident there. Cache files
    // may get evicted at any time, which then leads to renewed
    // downloading. See also "docs/fetcher.md" and
    // "docs/fetcher-cache-internals.md".
    optional bool cache = 4;

    // The fetcher's default behavior is to use the URI string's basename to
    // name the local copy. If this field is provided, the local copy will be
    // named with its value instead. If there is a directory component (which
    // must be a relative path), the local copy will be stored in that
    // subdirectory inside the sandbox.
    optional string output_file = 5;
  }

  repeated URI uris = 1;

  optional Environment environment = 2;

  // There are two ways to specify the command:
  // 1) If 'shell == true', the command will be launched via shell
  //		(i.e., /bin/sh -c 'value'). The 'value' specified will be
  //		treated as the shell command. The 'arguments' will be ignored.
  // 2) If 'shell == false', the command will be launched by passing
  //		arguments to an executable. The 'value' specified will be
  //		treated as the filename of the executable. The 'arguments'
  //		will be treated as the arguments to the executable. This is
  //		similar to how POSIX exec families launch processes (i.e.,
  //		execlp(value, arguments(0), arguments(1), ...)).
  // NOTE: The field 'value' is changed from 'required' to 'optional'
  // in 0.20.0. It will only cause issues if a new framework is
  // connecting to an old master.
  optional bool shell = 6 [default = true];
  optional string value = 3;
  repeated string arguments = 7;

  // Enables executor and tasks to run as a specific user. If the user
  // field is present both in FrameworkInfo and here, the CommandInfo
  // user value takes precedence.
  optional string user = 5;
}


/**
 * Describes information about an executor.
 */
message ExecutorInfo {
  enum Type {
    UNKNOWN = 0;

    // Mesos provides a simple built-in default executor that frameworks can
    // leverage to run shell commands and containers.
    //
    // NOTES:
    //
    // 1) `command` must not be set when using a default executor.
    //
    // 2) If "cpu" or "mem" resources are not set, default values of 0.1 cpu
    //    and 32 MB memory will be used; schedulers should account for these
    //    resources when accepting offers.
    //
    // 3) Default executor only accepts a *single* `LAUNCH` or `LAUNCH_GROUP`
    //    offer operation.
    DEFAULT = 1;

    // For frameworks that need custom functionality to run tasks, a `CUSTOM`
    // executor can be used. Note that `command` must be set when using a
    // `CUSTOM` executor.
    CUSTOM = 2;
  }

  // For backwards compatibility, if this field is not set when using `LAUNCH`
  // offer operation, Mesos will infer the type by checking if `command` is
  // set (`CUSTOM`) or unset (`DEFAULT`). `type` must be set when using
  // `LAUNCH_GROUP` offer operation.
  //
  // TODO(vinod): Add support for explicitly setting `type` to `DEFAULT `
  // in `LAUNCH` offer operation.
  optional Type type = 15;

  required ExecutorID executor_id = 1;
  optional FrameworkID framework_id = 8; // TODO(benh): Make this required.
  optional CommandInfo command = 7;

  // Executor provided with a container will launch the container
  // with the executor's CommandInfo and we expect the container to
  // act as a Mesos executor.
  optional ContainerInfo container = 11;

  repeated Resource resources = 5;
  optional string name = 9;

  // 'source' is an identifier style string used by frameworks to
  // track the source of an executor. This is useful when it's
  // possible for different executor ids to be related semantically.
  //
  // NOTE: 'source' is exposed alongside the resource usage of the
  // executor via JSON on the slave. This allows users to import usage
  // information into a time series database for monitoring.
  //
  // This field is deprecated since 1.0. Please use labels for
  // free-form metadata instead.
  optional string source = 10 [deprecated = true]; // Since 1.0.

  // This field can be used to pass arbitrary bytes to an executor.
  optional bytes data = 4;

  // Service discovery information for the executor. It is not
  // interpreted or acted upon by Mesos. It is up to a service
  // discovery system to use this information as needed and to handle
  // executors without service discovery information.
  optional DiscoveryInfo discovery = 12;

  // When shutting down an executor the agent will wait in a
  // best-effort manner for the grace period specified here
  // before forcibly destroying the container. The executor
  // must not assume that it will always be allotted the full
  // grace period, as the agent may decide to allot a shorter
  // period and failures / forcible terminations may occur.
  optional DurationInfo shutdown_grace_period = 13;

  // Labels are free-form key value pairs which are exposed through
  // master and slave endpoints. Labels will not be interpreted or
  // acted upon by Mesos itself. As opposed to the data field, labels
  // will be kept in memory on master and slave processes. Therefore,
  // labels should be used to tag executors with lightweight metadata.
  // Labels should not contain duplicate key-value pairs.
  optional Labels labels = 14;
}


/**
 * Describes a master. This will probably have more fields in the
 * future which might be used, for example, to link a framework webui
 * to a master webui.
 */
message MasterInfo {
  required string id = 1;

  // The IP address (only IPv4) as a packed 4-bytes integer,
  // stored in network order.  Deprecated, use `address.ip` instead.
  required uint32 ip = 2;

  // The TCP port the Master is listening on for incoming
  // HTTP requests; deprecated, use `address.port` instead.
  required uint32 port = 3 [default = 5050];

  // In the default implementation, this will contain information
  // about both the IP address, port and Master name; it should really
  // not be relied upon by external tooling/frameworks and be
  // considered an "internal" implementation field.
  optional string pid = 4;

  // The server's hostname, if available; it may be unreliable
  // in environments where the DNS configuration does not resolve
  // internal hostnames (eg, some public cloud providers).
  // Deprecated, use `address.hostname` instead.
  optional string hostname = 5;

  // The running Master version, as a string; taken from the
  // generated "master/version.hpp".
  optional string version = 6;

  // The full IP address (supports both IPv4 and IPv6 formats)
  // and supersedes the use of `ip`, `port` and `hostname`.
  // Since Mesos 0.24.
  optional Address address = 7;
}


/**
 * Describes a slave. Note that the 'id' field is only available after
 * a slave is registered with the master, and is made available here
 * to facilitate re-registration.
 */
message SlaveInfo {
  required string hostname = 1;
  optional int32 port = 8 [default = 5051];
  repeated Resource resources = 3;
  repeated Attribute attributes = 5;
  optional SlaveID id = 6;

  // Slave checkpointing is always enabled in recent Mesos versions;
  // the value of this field is ignored.
  // TODO(joerg84): Remove checkpoint field after deprecation cycle starting
  // with 0.27 (MESOS-2317).
  optional bool checkpoint = 7 [default = false];
}


/**
 * Describes an Attribute or Resource "value". A value is described
 * using the standard protocol buffer "union" trick.
 */
message Value {
  enum Type {
    SCALAR = 0;
    RANGES = 1;
    SET = 2;
    TEXT = 3;
  }

  message Scalar {
    // Scalar values are represented using floating point. To reduce
    // the chance of unpredictable floating point behavior due to
    // roundoff error, Mesos only supports three decimal digits of
    // precision for scalar resource values. That is, floating point
    // values are converted to a fixed point format that supports
    // three decimal digits of precision, and then converted back to
    // floating point on output. Any additional precision in scalar
    // resource values is discarded (via rounding).
    required double value = 1;
  }

  message Range {
    required uint64 begin = 1;
    required uint64 end = 2;
  }

  message Ranges {
    repeated Range range = 1;
  }

  message Set {
    repeated string item = 1;
  }

  message Text {
    required string value = 1;
  }

  required Type type = 1;
  optional Scalar scalar = 2;
  optional Ranges ranges = 3;
  optional Set set = 4;
  optional Text text = 5;
}


/**
 * Describes an attribute that can be set on a machine. For now,
 * attributes and resources share the same "value" type, but this may
 * change in the future and attributes may only be string based.
 */
message Attribute {
  required string name = 1;
  required Value.Type type = 2;
  optional Value.Scalar scalar = 3;
  optional Value.Ranges ranges = 4;
  optional Value.Set set = 6;
  optional Value.Text text = 5;
}


/**
 * Describes a resource on a machine. The `name` field is a string
 * like "cpus" or "mem" that indicates which kind of resource this is;
 * the rest of the fields describe the properties of the resource. A
 * resource can take on one of three types: scalar (double), a list of
 * finite and discrete ranges (e.g., [1-10, 20-30]), or a set of
 * items. A resource is described using the standard protocol buffer
 * "union" trick.
 *
 * Note that "disk" and "mem" resources are scalar values expressed in
 * megabytes. Fractional "cpus" values are allowed (e.g., "0.5"),
 * which correspond to partial shares of a CPU.
 */
message Resource {
  required string name = 1;
  required Value.Type type = 2;
  optional Value.Scalar scalar = 3;
  optional Value.Ranges ranges = 4;
  optional Value.Set set = 5;

  // The role that this resource is reserved for. If "*", this indicates
  // that the resource is unreserved. Otherwise, the resource will only
  // be offered to frameworks that belong to this role.
  optional string role = 6 [default = "*"];

  message ReservationInfo {
    // Describes a dynamic reservation. A dynamic reservation is
    // acquired by an operator via the '/reserve' HTTP endpoint or by
    // a framework via the offer cycle by sending back an
    // 'Offer::Operation::Reserve' message.
    // NOTE: We currently do not allow frameworks with role "*" to
    // make dynamic reservations.

    // Indicates the principal, if any, of the framework or operator
    // that reserved this resource. If reserved by a framework, the
    // field should match the `FrameworkInfo.principal`. It is used in
    // conjunction with the `UnreserveResources` ACL to determine
    // whether the entity attempting to unreserve this resource is
    // permitted to do so.
    optional string principal = 1;

    // Labels are free-form key value pairs that can be used to
    // associate arbitrary metadata with a reserved resource.  For
    // example, frameworks can use labels to identify the intended
    // purpose for a portion of the resources the framework has
    // reserved at a given slave. Labels should not contain duplicate
    // key-value pairs.
    optional Labels labels = 2;
  }

  // If this is set, this resource was dynamically reserved by an
  // operator or a framework. Otherwise, this resource is either unreserved
  // or statically reserved by an operator via the --resources flag.
  optional ReservationInfo reservation = 8;

  message DiskInfo {
    // Describes a persistent disk volume.
    //
    // A persistent disk volume will not be automatically garbage
    // collected if the task/executor/slave terminates, but will be
    // re-offered to the framework(s) belonging to the 'role'.
    //
    // NOTE: Currently, we do not allow persistent disk volumes
    // without a reservation (i.e., 'role' cannot be '*').
    message Persistence {
      // A unique ID for the persistent disk volume. This ID must be
      // unique per role on each slave. Although it is possible to use
      // the same ID on different slaves in the cluster and to reuse
      // IDs after a volume with that ID has been destroyed, both
      // practices are discouraged.
      required string id = 1;

      // This field indicates the principal of the operator or
      // framework that created this volume. It is used in conjunction
      // with the "destroy" ACL to determine whether an entity
      // attempting to destroy the volume is permitted to do so.
      //
      // NOTE: This field should match the FrameworkInfo.principal of
      // the framework that created the volume.
      optional string principal = 2;
    }

    optional Persistence persistence = 1;

    // Describes how this disk resource will be mounted in the
    // container. If not set, the disk resource will be used as the
    // sandbox. Otherwise, it will be mounted according to the
    // 'container_path' inside 'volume'. The 'host_path' inside
    // 'volume' is ignored.
    // NOTE: If 'volume' is set but 'persistence' is not set, the
    // volume will be automatically garbage collected after
    // task/executor terminates. Currently, if 'persistence' is set,
    // 'volume' must be set.
    optional Volume volume = 2;

    // Describes where a disk originates from.
    // TODO(jmlvanre): Add support for BLOCK devices.
    message Source {
      enum Type {
        PATH = 1;
        MOUNT = 2;
      }

      // A folder that can be located on a separate disk device. This
      // can be shared and carved up as necessary between frameworks.
      message Path {
        // Path to the folder (e.g., /mnt/raid/disk0).
        required string root = 1;
      }

      // A mounted file-system set up by the Agent administrator. This
      // can only be used exclusively: a framework can not accept a
      // partial amount of this disk.
      message Mount {
        // Path to mount point (e.g., /mnt/raid/disk0).
        required string root = 1;
      }

      required Type type = 1;
      optional Path path = 2;
      optional Mount mount = 3;
    }

    optional Source source = 3;
  }

  optional DiskInfo disk = 7;

  message RevocableInfo {}

  // If this is set, the resources are revocable, i.e., any tasks or
  // executors launched using these resources could get preempted or
  // throttled at any time. This could be used by frameworks to run
  // best effort tasks that do not need strict uptime or performance
  // guarantees. Note that if this is set, 'disk' or 'reservation'
  // cannot be set.
  optional RevocableInfo revocable = 9;

  // Allow the resource to be shared across tasks.
  message SharedInfo {}

  // If this is set, the resources are shared, i.e. multiple tasks
  // can be launched using this resource and all of them shall refer
  // to the same physical resource on the cluster. Note that only
  // persistent volumes can be shared currently.
  optional SharedInfo shared = 10;
}

/**
 * When the network bandwidth caps are enabled and the container
 * is over its limit, outbound packets may be either delayed or
 * dropped completely either because it exceeds the maximum bandwidth
 * allocation for a single container (the cap) or because the combined
 * network traffic of multiple containers on the host exceeds the
 * transmit capacity of the host (the share). We can report the
 * following statistics for each of these conditions exported directly
 * from the Linux Traffic Control Queueing Discipline.
 *
 * id         : name of the limiter, e.g. 'tx_bw_cap'
 * backlog    : number of packets currently delayed
 * bytes      : total bytes seen
 * drops      : number of packets dropped in total
 * overlimits : number of packets which exceeded allocation
 * packets    : total packets seen
 * qlen       : number of packets currently queued
 * rate_bps   : throughput in bytes/sec
 * rate_pps   : throughput in packets/sec
 * requeues   : number of times a packet has been delayed due to
 *              locking or device contention issues
 *
 * More information on the operation of Linux Traffic Control can be
 * found at http://www.lartc.org/lartc.html.
 */
message TrafficControlStatistics {
  required string id = 1;
  optional uint64 backlog = 2;
  optional uint64 bytes = 3;
  optional uint64 drops = 4;
  optional uint64 overlimits = 5;
  optional uint64 packets = 6;
  optional uint64 qlen = 7;
  optional uint64 ratebps = 8;
  optional uint64 ratepps = 9;
  optional uint64 requeues = 10;
}


message IpStatistics {
  optional int64 Forwarding =  1;
  optional int64 DefaultTTL =  2;
  optional int64 InReceives =  3;
  optional int64 InHdrErrors =  4;
  optional int64 InAddrErrors =  5;
  optional int64 ForwDatagrams =  6;
  optional int64 InUnknownProtos = 7;
  optional int64 InDiscards =  8;
  optional int64 InDelivers  = 9;
  optional int64 OutRequests = 10;
  optional int64 OutDiscards = 11;
  optional int64 OutNoRoutes = 12;
  optional int64 ReasmTimeout = 13;
  optional int64 ReasmReqds = 14;
  optional int64 ReasmOKs = 15;
  optional int64 ReasmFails = 16;
  optional int64 FragOKs = 17;
  optional int64 FragFails = 18;
  optional int64 FragCreates = 19;
}


message IcmpStatistics {
  optional int64 InMsgs = 1;
  optional int64 InErrors =  2;
  optional int64 InCsumErrors = 3;
  optional int64 InDestUnreachs =  4;
  optional int64 InTimeExcds = 5;
  optional int64 InParmProbs =  6;
  optional int64 InSrcQuenchs = 7;
  optional int64 InRedirects = 8;
  optional int64 InEchos = 9;
  optional int64 InEchoReps = 10;
  optional int64 InTimestamps = 11;
  optional int64 InTimestampReps = 12;
  optional int64 InAddrMasks = 13;
  optional int64 InAddrMaskReps = 14;
  optional int64 OutMsgs = 15;
  optional int64 OutErrors = 16;
  optional int64 OutDestUnreachs = 17;
  optional int64 OutTimeExcds = 18;
  optional int64 OutParmProbs = 19;
  optional int64 OutSrcQuenchs = 20;
  optional int64 OutRedirects = 21;
  optional int64 OutEchos = 22;
  optional int64 OutEchoReps = 23;
  optional int64 OutTimestamps = 24;
  optional int64 OutTimestampReps = 25;
  optional int64 OutAddrMasks = 26;
  optional int64 OutAddrMaskReps = 27;
}


message TcpStatistics {
  optional int64 RtoAlgorithm = 1;
  optional int64 RtoMin = 2;
  optional int64 RtoMax = 3;
  optional int64 MaxConn = 4;
  optional int64 ActiveOpens = 5;
  optional int64 PassiveOpens = 6;
  optional int64 AttemptFails = 7;
  optional int64 EstabResets = 8;
  optional int64 CurrEstab = 9;
  optional int64 InSegs = 10;
  optional int64 OutSegs = 11;
  optional int64 RetransSegs = 12;
  optional int64 InErrs = 13;
  optional int64 OutRsts = 14;
  optional int64 InCsumErrors = 15;
}


message UdpStatistics {
  optional int64 InDatagrams = 1;
  optional int64 NoPorts = 2;
  optional int64 InErrors = 3;
  optional int64 OutDatagrams = 4;
  optional int64 RcvbufErrors = 5;
  optional int64 SndbufErrors = 6;
  optional int64 InCsumErrors = 7;
  optional int64 IgnoredMulti = 8;
}


message SNMPStatistics {
  optional IpStatistics ip_stats = 1;
  optional IcmpStatistics icmp_stats = 2;
  optional TcpStatistics tcp_stats = 3;
  optional UdpStatistics udp_stats = 4;
}


/**
 * A snapshot of resource usage statistics.
 */
message ResourceStatistics {
  required double timestamp = 1; // Snapshot time, in seconds since the Epoch.

  optional uint32 processes = 30;
  optional uint32 threads = 31;

  // CPU Usage Information:
  // Total CPU time spent in user mode, and kernel mode.
  optional double cpus_user_time_secs = 2;
  optional double cpus_system_time_secs = 3;

  // Number of CPUs allocated.
  optional double cpus_limit = 4;

  // cpu.stat on process throttling (for contention issues).
  optional uint32 cpus_nr_periods = 7;
  optional uint32 cpus_nr_throttled = 8;
  optional double cpus_throttled_time_secs = 9;

  // Memory Usage Information:

  // mem_total_bytes was added in 0.23.0 to represent the total memory
  // of a process in RAM (as opposed to in Swap). This was previously
  // reported as mem_rss_bytes, which was also changed in 0.23.0 to
  // represent only the anonymous memory usage, to keep in sync with
  // Linux kernel's (arguably erroneous) use of terminology.
  optional uint64 mem_total_bytes = 36;

  // Total memory + swap usage. This is set if swap is enabled.
  optional uint64 mem_total_memsw_bytes = 37;

  // Hard memory limit for a container.
  optional uint64 mem_limit_bytes = 6;

  // Soft memory limit for a container.
  optional uint64 mem_soft_limit_bytes = 38;

  // Broken out memory usage information: pagecache, rss (anonymous),
  // mmaped files and swap.

  // TODO(chzhcn) mem_file_bytes and mem_anon_bytes are deprecated in
  // 0.23.0 and will be removed in 0.24.0.
  optional uint64 mem_file_bytes = 10;
  optional uint64 mem_anon_bytes = 11;

  // mem_cache_bytes is added in 0.23.0 to represent page cache usage.
  optional uint64 mem_cache_bytes = 39;

  // Since 0.23.0, mem_rss_bytes is changed to represent only
  // anonymous memory usage. Note that neither its requiredness, type,
  // name nor numeric tag has been changed.
  optional uint64 mem_rss_bytes = 5;

  optional uint64 mem_mapped_file_bytes = 12;
  // This is only set if swap is enabled.
  optional uint64 mem_swap_bytes = 40;
  optional uint64 mem_unevictable_bytes = 41;

  // Number of occurrences of different levels of memory pressure
  // events reported by memory cgroup. Pressure listening (re)starts
  // with these values set to 0 when slave (re)starts. See
  // https://www.kernel.org/doc/Documentation/cgroups/memory.txt for
  // more details.
  optional uint64 mem_low_pressure_counter = 32;
  optional uint64 mem_medium_pressure_counter = 33;
  optional uint64 mem_critical_pressure_counter = 34;

  // Disk Usage Information for executor working directory.
  optional uint64 disk_limit_bytes = 26;
  optional uint64 disk_used_bytes = 27;

  // Perf statistics.
  optional PerfStatistics perf = 13;

  // Network Usage Information:
  optional uint64 net_rx_packets = 14;
  optional uint64 net_rx_bytes = 15;
  optional uint64 net_rx_errors = 16;
  optional uint64 net_rx_dropped = 17;
  optional uint64 net_tx_packets = 18;
  optional uint64 net_tx_bytes = 19;
  optional uint64 net_tx_errors = 20;
  optional uint64 net_tx_dropped = 21;

  // The kernel keeps track of RTT (round-trip time) for its TCP
  // sockets. RTT is a way to tell the latency of a container.
  optional double net_tcp_rtt_microsecs_p50 = 22;
  optional double net_tcp_rtt_microsecs_p90 = 23;
  optional double net_tcp_rtt_microsecs_p95 = 24;
  optional double net_tcp_rtt_microsecs_p99 = 25;

  optional double net_tcp_active_connections = 28;
  optional double net_tcp_time_wait_connections = 29;

  // Network traffic flowing into or out of a container can be delayed
  // or dropped due to congestion or policy inside and outside the
  // container.
  repeated TrafficControlStatistics net_traffic_control_statistics = 35;

  // Network SNMP statistics for each container.
  optional SNMPStatistics net_snmp_statistics = 42;
}


/**
 * Describes a snapshot of the resource usage for executors.
 */
message ResourceUsage {
  message Executor {
    required ExecutorInfo executor_info = 1;

    // This includes resources used by the executor itself
    // as well as its active tasks.
    repeated Resource allocated = 2;

    // Current resource usage. If absent, the containerizer
    // cannot provide resource usage.
    optional ResourceStatistics statistics = 3;

    // The container id for the executor specified in the executor_info field.
    required ContainerID container_id = 4;

    message Task {
      required string name = 1;
      required TaskID id = 2;
      repeated Resource resources = 3;
      optional Labels labels = 4;
    }

    // Non-terminal tasks.
    repeated Task tasks = 5;
  }

  repeated Executor executors = 1;

  // Slave's total resources including checkpointed dynamic
  // reservations and persistent volumes.
  repeated Resource total = 2;
}


/**
 * Describes a sample of events from "perf stat". Only available on
 * Linux.
 *
 * NOTE: Each optional field matches the name of a perf event (see
 * "perf list") with the following changes:
 * 1. Names are downcased.
 * 2. Hyphens ('-') are replaced with underscores ('_').
 * 3. Events with alternate names use the name "perf stat" returns,
 *    e.g., for the event "cycles OR cpu-cycles" perf always returns
 *    cycles.
 */
message PerfStatistics {
  required double timestamp = 1; // Start of sample interval, in seconds since the Epoch.
  required double duration = 2;  // Duration of sample interval, in seconds.

  // Hardware event.
  optional uint64 cycles = 3;
  optional uint64 stalled_cycles_frontend = 4;
  optional uint64 stalled_cycles_backend = 5;
  optional uint64 instructions = 6;
  optional uint64 cache_references = 7;
  optional uint64 cache_misses = 8;
  optional uint64 branches = 9;
  optional uint64 branch_misses = 10;
  optional uint64 bus_cycles = 11;
  optional uint64 ref_cycles = 12;

  // Software event.
  optional double cpu_clock = 13;
  optional double task_clock = 14;
  optional uint64 page_faults = 15;
  optional uint64 minor_faults = 16;
  optional uint64 major_faults = 17;
  optional uint64 context_switches = 18;
  optional uint64 cpu_migrations = 19;
  optional uint64 alignment_faults = 20;
  optional uint64 emulation_faults = 21;

  // Hardware cache event.
  optional uint64 l1_dcache_loads = 22;
  optional uint64 l1_dcache_load_misses = 23;
  optional uint64 l1_dcache_stores = 24;
  optional uint64 l1_dcache_store_misses = 25;
  optional uint64 l1_dcache_prefetches = 26;
  optional uint64 l1_dcache_prefetch_misses = 27;
  optional uint64 l1_icache_loads = 28;
  optional uint64 l1_icache_load_misses = 29;
  optional uint64 l1_icache_prefetches = 30;
  optional uint64 l1_icache_prefetch_misses = 31;
  optional uint64 llc_loads = 32;
  optional uint64 llc_load_misses = 33;
  optional uint64 llc_stores = 34;
  optional uint64 llc_store_misses = 35;
  optional uint64 llc_prefetches = 36;
  optional uint64 llc_prefetch_misses = 37;
  optional uint64 dtlb_loads = 38;
  optional uint64 dtlb_load_misses = 39;
  optional uint64 dtlb_stores = 40;
  optional uint64 dtlb_store_misses = 41;
  optional uint64 dtlb_prefetches = 42;
  optional uint64 dtlb_prefetch_misses = 43;
  optional uint64 itlb_loads = 44;
  optional uint64 itlb_load_misses = 45;
  optional uint64 branch_loads = 46;
  optional uint64 branch_load_misses = 47;
  optional uint64 node_loads = 48;
  optional uint64 node_load_misses = 49;
  optional uint64 node_stores = 50;
  optional uint64 node_store_misses = 51;
  optional uint64 node_prefetches = 52;
  optional uint64 node_prefetch_misses = 53;
}


/**
 * Describes a request for resources that can be used by a framework
 * to proactively influence the allocator.  If 'slave_id' is provided
 * then this request is assumed to only apply to resources on that
 * slave.
 */
message Request {
  optional SlaveID slave_id = 1;
  repeated Resource resources = 2;
}


/**
 * Describes some resources available on a slave. An offer only
 * contains resources from a single slave.
 */
message Offer {
  required OfferID id = 1;
  required FrameworkID framework_id = 2;
  required SlaveID slave_id = 3;
  required string hostname = 4;

  // URL for reaching the slave running on the host.
  optional URL url = 8;

  repeated Resource resources = 5;
  repeated Attribute attributes = 7;
  repeated ExecutorID executor_ids = 6;

  // Signifies that the resources in this Offer may be unavailable during
  // the given interval.  Any tasks launched using these resources may be
  // killed when the interval arrives.  For example, these resources may be
  // part of a planned maintenance schedule.
  //
  // This field only provides information about a planned unavailability.
  // The unavailability interval may not necessarily start at exactly this
  // interval, nor last for exactly the duration of this interval.
  // The unavailability may also be forever!  See comments in
  // `Unavailability` for more details.
  optional Unavailability unavailability = 9;

  // Defines an operation that can be performed against offers.
  message Operation {
    enum Type {
      UNKNOWN = 0;
      LAUNCH = 1;
      LAUNCH_GROUP = 6;
      RESERVE = 2;
      UNRESERVE = 3;
      CREATE = 4;
      DESTROY = 5;
    }

    // TODO(vinod): Deprecate this in favor of `LaunchGroup` below.
    message Launch {
      repeated TaskInfo task_infos = 1;
    }

    // Unlike `Launch` above, all the tasks in a `task_group` are
    // atomically delivered to an executor.
    //
    // `NetworkInfo` set on executor will be shared by all tasks in
    // the task group.
    //
    // TODO(vinod): Any volumes set on executor could be used by a
    // task by explicitly setting `Volume.source` in its resources.
    message LaunchGroup {
      required ExecutorInfo executor = 1;
      required TaskGroupInfo task_group = 2;
    }

    message Reserve {
      repeated Resource resources = 1;
    }

    message Unreserve {
      repeated Resource resources = 1;
    }

    message Create {
      repeated Resource volumes = 1;
    }

    message Destroy {
      repeated Resource volumes = 1;
    }

    optional Type type = 1;
    optional Launch launch = 2;
    optional LaunchGroup launch_group = 7;
    optional Reserve reserve = 3;
    optional Unreserve unreserve = 4;
    optional Create create = 5;
    optional Destroy destroy = 6;
  }
}


/**
 * A request to return some resources occupied by a framework.
 */
message InverseOffer {
  // This is the same OfferID as found in normal offers, which allows
  // re-use of some of the OfferID-only messages.
  required OfferID id = 1;

  // URL for reaching the slave running on the host.  This enables some
  // optimizations as described in MESOS-3012, such as allowing the
  // scheduler driver to bypass the master and talk directly with a slave.
  optional URL url = 2;

  // The framework that should release its resources.
  // If no specifics are provided (i.e. which slave), all the framework's
  // resources are requested back.
  required FrameworkID framework_id = 3;

  // Specified if the resources need to be released from a particular slave.
  // All the framework's resources on this slave are requested back,
  // unless further qualified by the `resources` field.
  optional SlaveID slave_id = 4;

  // This InverseOffer represents a planned unavailability event in the
  // specified interval.  Any tasks running on the given framework or slave
  // may be killed when the interval arrives.  Therefore, frameworks should
  // aim to gracefully terminate tasks prior to the arrival of the interval.
  //
  // For reserved resources, the resources are expected to be returned to the
  // framework after the unavailability interval.  This is an expectation,
  // not a guarantee.  For example, if the unavailability duration is not set,
  // the resources may be removed permanently.
  //
  // For other resources, there is no guarantee that requested resources will
  // be returned after the unavailability interval.  The allocator has no
  // obligation to re-offer these resources to the prior framework after
  // the unavailability.
  required Unavailability unavailability = 5;

  // A list of resources being requested back from the framework,
  // on the slave identified by `slave_id`.  If no resources are specified
  // then all resources are being requested back.  For the purpose of
  // maintenance, this field is always empty (maintenance always requests
  // all resources back).
  repeated Resource resources = 6;

  // TODO(josephw): Add additional options for narrowing down the resources
  // being requested back.  Such as specific executors, tasks, etc.
}


/**
 * Describes a task. Passed from the scheduler all the way to an
 * executor (see SchedulerDriver::launchTasks and
 * Executor::launchTask). Either ExecutorInfo or CommandInfo should be set.
 * A different executor can be used to launch this task, and subsequent tasks
 * meant for the same executor can reuse the same ExecutorInfo struct.
 */
message TaskInfo {
  required string name = 1;
  required TaskID task_id = 2;
  required SlaveID slave_id = 3;
  repeated Resource resources = 4;
  optional ExecutorInfo executor = 5;
  optional CommandInfo command = 7;

  // Task provided with a container will launch the container as part
  // of this task paired with the task's CommandInfo.
  optional ContainerInfo container = 9;

  // A health check for the task. Implemented for executor-less
  // command-based tasks. For tasks that specify an executor, it is
  // the executor's responsibility to implement the health checking.
  optional HealthCheck health_check = 8;

  // A kill policy for the task. Implemented for executor-less
  // command-based and docker tasks. For tasks that specify other
  // executor, it is the executor's responsibility to implement
  // the kill policy.
  optional KillPolicy kill_policy = 12;

  optional bytes data = 6;

  // Labels are free-form key value pairs which are exposed through
  // master and slave endpoints. Labels will not be interpreted or
  // acted upon by Mesos itself. As opposed to the data field, labels
  // will be kept in memory on master and slave processes. Therefore,
  // labels should be used to tag tasks with light-weight meta-data.
  // Labels should not contain duplicate key-value pairs.
  optional Labels labels = 10;

  // Service discovery information for the task. It is not interpreted
  // or acted upon by Mesos. It is up to a service discovery system
  // to use this information as needed and to handle tasks without
  // service discovery information.
  optional DiscoveryInfo discovery = 11;
}


/**
 * Describes a group of tasks that belong to an executor. The
 * executor will receive the task group in a single message to
 * allow the group to be launched "atomically".
 *
 * NOTES:
 * 1) `TaskInfo.executor` must not be set.
 * 2) `NetworkInfo` must not be set inside task's `ContainerInfo`.
 */
message TaskGroupInfo {
  repeated TaskInfo tasks = 1;
}


// TODO(bmahler): Add executor_uuid here, and send it to the master. This will
// allow us to expose executor work directories for tasks in the webui when
// looking from the master level. Currently only the slave knows which run the
// task belongs to.
/**
 * Describes a task, similar to `TaskInfo`.
 *
 * `Task` is used in some of the Mesos messages found below.
 * `Task` is used instead of `TaskInfo` if:
 *   1) we need additional IDs, such as a specific
 *      framework, executor, or agent; or
 *   2) we do not need the additional data, such as the command run by the
 *      task or the health checks.  These additional fields may be large and
 *      unnecessary for some Mesos messages.
 *
 * `Task` is generally constructed from a `TaskInfo`.  See protobuf::createTask.
 */
message Task {
  required string name = 1;
  required TaskID task_id = 2;
  required FrameworkID framework_id = 3;
  optional ExecutorID executor_id = 4;
  required SlaveID slave_id = 5;
  required TaskState state = 6; // Latest state of the task.
  repeated Resource resources = 7;
  repeated TaskStatus statuses = 8;

  // These fields correspond to the state and uuid of the latest
  // status update forwarded to the master.
  // NOTE: Either both the fields must be set or both must be unset.
  optional TaskState status_update_state = 9;
  optional bytes status_update_uuid = 10;

  optional Labels labels = 11;

  // Service discovery information for the task. It is not interpreted
  // or acted upon by Mesos. It is up to a service discovery system
  // to use this information as needed and to handle tasks without
  // service discovery information.
  optional DiscoveryInfo discovery = 12;

  // Container information for the task.
  optional ContainerInfo container = 13;

  // Specific user under which task is running.
  optional string user = 14;
}


/**
 * Describes possible task states. IMPORTANT: Mesos assumes tasks that
 * enter terminal states (see below) imply the task is no longer
 * running and thus clean up any thing associated with the task
 * (ultimately offering any resources being consumed by that task to
 * another task).
 */
enum TaskState {
  TASK_STAGING = 6;  // Initial state. Framework status updates should not use.
  TASK_STARTING = 0; // The task is being launched by the executor.
  TASK_RUNNING = 1;

  // NOTE: This should only be sent when the framework has
  // the TASK_KILLING_STATE capability.
  TASK_KILLING = 8;  // The task is being killed by the executor.

  TASK_FINISHED = 2; // TERMINAL: The task finished successfully.
  TASK_FAILED = 3;   // TERMINAL: The task failed to finish successfully.
  TASK_KILLED = 4;   // TERMINAL: The task was killed by the executor.
  TASK_ERROR = 7;    // TERMINAL: The task description contains an error.

  // This is only sent when the framework does NOT opt-in to the
  // PARTITION_AWARE capability.
  TASK_LOST = 5;     // TERMINAL: The task failed but can be rescheduled.

  // The following task statuses are only sent when the framework
  // opts-in to the PARTITION_AWARE capability.

  // The task failed to launch because of a transient error. The
  // task's executor never started running. Unlike TASK_ERROR, the
  // task description is valid -- attempting to launch the task again
  // may be successful. This is a terminal state.
  TASK_DROPPED = 9;

  // The task was running on an agent that has lost contact with the
  // master, typically due to a network failure or partition. The task
  // may or may not still be running.
  TASK_UNREACHABLE = 10;

  // The task was running on an agent that has been shutdown (e.g.,
  // the agent become partitioned, rebooted, and then reconnected to
  // the master; any tasks running before the reboot will transition
  // from UNREACHABLE to GONE). The task is no longer running. This is
  // a terminal state.
  TASK_GONE = 11;

  // The task was running on an agent that the master cannot contact;
  // the operator has asserted that the agent has been shutdown, but
  // this has not been directly confirmed by the master. If the
  // operator is correct, the task is not running and this is a
  // terminal state; if the operator is mistaken, the task might still
  // be running, and might return to the RUNNING state in the future.
  TASK_GONE_BY_OPERATOR = 12;

  // The master has no knowledge of the task. This is typically
  // because either (a) the master never had knowledge of the task, or
  // (b) the master forgot about the task because it garbaged
  // collected its metadata about the task. The task may or may not
  // still be running.
  TASK_UNKNOWN = 13;
}


/**
 * Describes the current status of a task.
 */
message TaskStatus {
  // Describes the source of the task status update.
  enum Source {
    SOURCE_MASTER = 0;
    SOURCE_SLAVE = 1;
    SOURCE_EXECUTOR = 2;
  }

  // Detailed reason for the task status update.
  //
  // TODO(bmahler): Differentiate between slave removal reasons
  // (e.g. unhealthy vs. unregistered for maintenance).
  enum Reason {
    // TODO(jieyu): The default value when a caller doesn't check for
    // presence is 0 and so ideally the 0 reason is not a valid one.
    // Since this is not used anywhere, consider removing this reason.
    REASON_COMMAND_EXECUTOR_FAILED = 0;

    REASON_CONTAINER_LAUNCH_FAILED = 21;
    REASON_CONTAINER_LIMITATION = 19;
    REASON_CONTAINER_LIMITATION_DISK = 20;
    REASON_CONTAINER_LIMITATION_MEMORY = 8;
    REASON_CONTAINER_PREEMPTED = 17;
    REASON_CONTAINER_UPDATE_FAILED = 22;
    REASON_EXECUTOR_REGISTRATION_TIMEOUT = 23;
    REASON_EXECUTOR_REREGISTRATION_TIMEOUT = 24;
    REASON_EXECUTOR_TERMINATED = 1;
    REASON_EXECUTOR_UNREGISTERED = 2;
    REASON_FRAMEWORK_REMOVED = 3;
    REASON_GC_ERROR = 4;
    REASON_INVALID_FRAMEWORKID = 5;
    REASON_INVALID_OFFERS = 6;
    REASON_MASTER_DISCONNECTED = 7;
    REASON_RECONCILIATION = 9;
    REASON_RESOURCES_UNKNOWN = 18;
    REASON_SLAVE_DISCONNECTED = 10;
    REASON_SLAVE_REMOVED = 11;
    REASON_SLAVE_RESTARTED = 12;
    REASON_SLAVE_UNKNOWN = 13;
    REASON_TASK_GROUP_INVALID = 25;
    REASON_TASK_GROUP_UNAUTHORIZED = 26;
    REASON_TASK_INVALID = 14;
    REASON_TASK_UNAUTHORIZED = 15;
    REASON_TASK_UNKNOWN = 16;
  }

  required TaskID task_id = 1;
  required TaskState state = 2;
  optional string message = 4; // Possible message explaining state.
  optional Source source = 9;
  optional Reason reason = 10;
  optional bytes data = 3;
  optional SlaveID slave_id = 5;
  optional ExecutorID executor_id = 7; // TODO(benh): Use in master/slave.
  optional double timestamp = 6;

  // Statuses that are delivered reliably to the scheduler will
  // include a 'uuid'. The status is considered delivered once
  // it is acknowledged by the scheduler. Schedulers can choose
  // to either explicitly acknowledge statuses or let the scheduler
  // driver implicitly acknowledge (default).
  //
  // TODO(bmahler): This is currently overwritten in the scheduler
  // driver and executor driver, but executors will need to set this
  // to a valid RFC-4122 UUID if using the HTTP API.
  optional bytes uuid = 11;

  // Describes whether the task has been determined to be healthy
  // (true) or unhealthy (false) according to the HealthCheck field in
  // the command info.
  optional bool healthy = 8;

  // Labels are free-form key value pairs which are exposed through
  // master and slave endpoints. Labels will not be interpreted or
  // acted upon by Mesos itself. As opposed to the data field, labels
  // will be kept in memory on master and slave processes. Therefore,
  // labels should be used to tag TaskStatus message with light-weight
  // meta-data. Labels should not contain duplicate key-value pairs.
  optional Labels labels = 12;

  // Container related information that is resolved dynamically such as
  // network address.
  optional ContainerStatus container_status = 13;
}


/**
 * Describes possible filters that can be applied to unused resources
 * (see SchedulerDriver::launchTasks) to influence the allocator.
 */
message Filters {
  // Time to consider unused resources refused. Note that all unused
  // resources will be considered refused and use the default value
  // (below) regardless of whether Filters was passed to
  // SchedulerDriver::launchTasks. You MUST pass Filters with this
  // field set to change this behavior (i.e., get another offer which
  // includes unused resources sooner or later than the default).
  optional double refuse_seconds = 1 [default = 5.0];
}


/**
* Describes a collection of environment variables. This is used with
* CommandInfo in order to set environment variables before running a
* command.
*/
message Environment {
  message Variable {
    required string name = 1;
    required string value = 2;
  }

  repeated Variable variables = 1;
}


/**
 * A generic (key, value) pair used in various places for parameters.
 */
message Parameter {
  required string key = 1;
  required string value = 2;
}


/**
 * Collection of Parameter.
 */
message Parameters {
  repeated Parameter parameter = 1;
}


/**
 * Credential used in various places for authentication and
 * authorization.
 *
 * NOTE: A 'principal' is different from 'FrameworkInfo.user'. The
 * former is used for authentication and authorization while the
 * latter is used to determine the default user under which the
 * framework's executors/tasks are run.
 */
message Credential {
  required string principal = 1;
  optional string secret = 2;
}


/**
 * Credentials used for framework authentication, HTTP authentication
 * (where the common 'username' and 'password' are captured as
 * 'principal' and 'secret' respectively), etc.
 */
message Credentials {
  repeated Credential credentials = 1;
}


/**
 * Rate (queries per second, QPS) limit for messages from a framework to master.
 * Strictly speaking they are the combined rate from all frameworks of the same
 * principal.
 */
message RateLimit {
  // Leaving QPS unset gives it unlimited rate (i.e., not throttled),
  // which also implies unlimited capacity.
  optional double qps = 1;

  // Principal of framework(s) to be throttled. Should match
  // FrameworkInfo.principal and Credential.principal (if using authentication).
  required string principal = 2;

  // Max number of outstanding messages from frameworks of this principal
  // allowed by master before the next message is dropped and an error is sent
  // back to the sender. Messages received before the capacity is reached are
  // still going to be processed after the error is sent.
  // If unspecified, this principal is assigned unlimited capacity.
  // NOTE: This value is ignored if 'qps' is not set.
  optional uint64 capacity = 3;
}


/**
 * Collection of RateLimit.
 * Frameworks without rate limits defined here are not throttled unless
 * 'aggregate_default_qps' is specified.
 */
message RateLimits {
  // Items should have unique principals.
  repeated RateLimit limits = 1;

  // All the frameworks not specified in 'limits' get this default rate.
  // This rate is an aggregate rate for all of them, i.e., their combined
  // traffic is throttled together at this rate.
  optional double aggregate_default_qps = 2;

  // All the frameworks not specified in 'limits' get this default capacity.
  // This is an aggregate value similar to 'aggregate_default_qps'.
  optional uint64 aggregate_default_capacity = 3;
}


/**
 * Describe an image used by tasks or executors. Note that it's only
 * for tasks or executors launched by MesosContainerizer currently.
 */
message Image {
  enum Type {
    APPC = 1;
    DOCKER = 2;
  }

  // Protobuf for specifying an Appc container image. See:
  // https://github.com/appc/spec/blob/master/spec/aci.md
  message Appc {
    // The name of the image.
    required string name = 1;

    // An image ID is a string of the format "hash-value", where
    // "hash" is the hash algorithm used and "value" is the hex
    // encoded string of the digest. Currently the only permitted
    // hash algorithm is sha512.
    optional string id = 2;

    // Optional labels. Suggested labels: "version", "os", and "arch".
    optional Labels labels = 3;
  }

  message Docker {
    // The name of the image. Expected format:
    //   [REGISTRY_HOST[:REGISTRY_PORT]/]REPOSITORY[:TAG|@TYPE:DIGEST]
    //
    // See: https://docs.docker.com/reference/commandline/pull/
    required string name = 1;

    // Credential to authenticate with docker registry.
    // NOTE: This is not encrypted, therefore framework and operators
    // should enable SSL when passing this information.
    optional Credential credential = 2;
  }

  required Type type = 1;

  // Only one of the following image messages should be set to match
  // the type.
  optional Appc appc = 2;
  optional Docker docker = 3;

  // With this flag set to false, the mesos containerizer will pull
  // the docker/appc image from the registry even if the image is
  // already downloaded on the agent.
  optional bool cached = 4 [default = true];
}


/**
 * Describes a volume mapping either from host to container or vice
 * versa. Both paths can either refer to a directory or a file.
 */
message Volume {
  enum Mode {
    RW = 1; // read-write.
    RO = 2; // read-only.
  }

  // TODO(gyliu513): Make this as `optional` after deprecation cycle of 1.0.
  required Mode mode = 3;

  // Path pointing to a directory or file in the container. If the
  // path is a relative path, it is relative to the container work
  // directory. If the path is an absolute path, that path must
  // already exist.
  required string container_path = 1;

  // The following specifies the source of this volume. At most one of
  // the following should be set.

  // Absolute path pointing to a directory or file on the host or a
  // path relative to the container work directory.
  optional string host_path = 2;

  // The source of the volume is an Image which describes a root
  // filesystem which will be provisioned by Mesos.
  optional Image image = 4;

  // Describes where a volume originates from.
  message Source {
    enum Type {
      // This must be the first enum value in this list, to
      // ensure that if 'type' is not set, the default value
      // is UNKNOWN. This enables enum values to be added
      // in a backwards-compatible way. See: MESOS-4997.
      UNKNOWN = 0;

      // TODO(gyliu513): Add HOST_PATH and IMAGE as volume source type.
      DOCKER_VOLUME = 1;
    }

    message DockerVolume {
      // Driver of the volume, it can be flocker, convoy, raxrey etc.
      optional string driver = 1;

      // Name of the volume.
      required string name = 2;

      // Volume driver specific options.
      optional Parameters driver_options = 3;
    }

    // Enum fields should be optional, see: MESOS-4997.
    optional Type type = 1;

    // The following specifies the source of this volume. At most one of
    // the following should be set.

    // The source of the volume created by docker volume driver.
    optional DockerVolume docker_volume = 2;
  }

  optional Source source = 5;
}


/**
 * Describes a network request from a framework as well as network resolution
 * provided by Mesos.
 *
 * A framework may request the network isolator on the Agent to isolate the
 * container in a network namespace and create a virtual network interface.
 * The `NetworkInfo` message describes the properties of that virtual
 * interface, including the IP addresses and network isolation policy
 * (network group membership).
 *
 * The NetworkInfo message is not interpreted by the Master or Agent and is
 * intended to be used by Agent and Master modules implementing network
 * isolation. If the modules are missing, the message is simply ignored. In
 * future, the task launch will fail if there is no module providing the
 * network isolation capabilities (MESOS-3390).
 *
 * An executor, Agent, or an Agent module may append NetworkInfos inside
 * TaskStatus::container_status to provide information such as the container IP
 * address and isolation groups.
 */
message NetworkInfo {
  enum Protocol {
    IPv4 = 1;
    IPv6 = 2;
  }

  // Specifies a request for an IP address, or reports the assigned container
  // IP address.
  //
  // Users can request an automatically assigned IP (for example, via an
  // IPAM service) or a specific IP by adding a NetworkInfo to the
  // ContainerInfo for a task.  On a request, specifying neither `protocol`
  // nor `ip_address` means that any available address may be assigned.
  message IPAddress {
    // Specify IP address requirement. Set protocol to the desired value to
    // request the network isolator on the Agent to assign an IP address to the
    // container being launched. If a specific IP address is specified in
    // ip_address, this field should not be set.
    optional Protocol protocol = 1;

    // Statically assigned IP provided by the Framework. This IP will be
    // assigned to the container by the network isolator module on the Agent.
    // This field should not be used with the protocol field above.
    //
    // If an explicit address is requested but is unavailable, the network
    // isolator should fail the task.
    optional string ip_address = 2;
  }

  // When included in a ContainerInfo, each of these represent a
  // request for an IP address. Each request can specify an explicit address
  // or the IP protocol to use.
  //
  // When included in a TaskStatus message, these inform the framework
  // scheduler about the IP addresses that are bound to the container
  // interface. When there are no custom network isolator modules installed,
  // this field is filled in automatically with the Agent IP address.
  repeated IPAddress ip_addresses = 5;

  // Name of the network which will be used by network isolator to determine
  // the network that the container joins. It's up to the network isolator
  // to decide how to interpret this field.
  optional string name = 6;

  // A group is the name given to a set of logically-related interfaces that
  // are allowed to communicate among themselves. Network traffic is allowed
  // between two container interfaces that share at least one network group.
  // For example, one might want to create separate groups for isolating dev,
  // testing, qa and prod deployment environments.
  repeated string groups = 3;

  // To tag certain metadata to be used by Isolator/IPAM, e.g., rack, etc.
  optional Labels labels = 4;

  // Specifies a port mapping request for the task on this network.
  message PortMapping {
    required uint32 host_port = 1;
    required uint32 container_port = 2;
    // Protocol to expose as (ie: tcp, udp).
    optional string protocol = 3;
  }

  repeated PortMapping port_mappings = 7;
};


/**
 * Encapsulation of `Capabilities` supported by Linux.
 * Reference: http://linux.die.net/man/7/capabilities.
 */
message CapabilityInfo {
  // We start the actual values at an offset(1000) because Protobuf 2
  // uses the first value as the default one. Separating the default
  // value from the real first value helps to disambiguate them. This
  // is especially valuable for backward compatibility.
  // See: MESOS-4997.
  enum Capability {
    UNKNOWN = 0;
    CHOWN = 1000;
    DAC_OVERRIDE = 1001;
    DAC_READ_SEARCH = 1002;
    FOWNER = 1003;
    FSETID = 1004;
    KILL = 1005;
    SETGID = 1006;
    SETUID = 1007;
    SETPCAP = 1008;
    LINUX_IMMUTABLE = 1009;
    NET_BIND_SERVICE = 1010;
    NET_BROADCAST = 1011;
    NET_ADMIN = 1012;
    NET_RAW = 1013;
    IPC_LOCK = 1014;
    IPC_OWNER = 1015;
    SYS_MODULE = 1016;
    SYS_RAWIO = 1017;
    SYS_CHROOT = 1018;
    SYS_PTRACE = 1019;
    SYS_PACCT = 1020;
    SYS_ADMIN = 1021;
    SYS_BOOT = 1022;
    SYS_NICE = 1023;
    SYS_RESOURCE = 1024;
    SYS_TIME = 1025;
    SYS_TTY_CONFIG = 1026;
    MKNOD = 1027;
    LEASE = 1028;
    AUDIT_WRITE = 1029;
    AUDIT_CONTROL = 1030;
    SETFCAP = 1031;
    MAC_OVERRIDE = 1032;
    MAC_ADMIN = 1033;
    SYSLOG = 1034;
    WAKE_ALARM = 1035;
    BLOCK_SUSPEND = 1036;
    AUDIT_READ = 1037;
  }

  repeated Capability capabilities = 1;
}


/**
 * Encapsulation for Linux specific configuration.
 * E.g, capabilities, limits etc.
 */
message LinuxInfo {
  // Represents the capability whitelist.
  optional CapabilityInfo capability_info = 1;
}


/**
 * Describes a container configuration and allows extensible
 * configurations for different container implementations.
 */
message ContainerInfo {
  // All container implementation types.
  enum Type {
    DOCKER = 1;
    MESOS = 2;
  }

  message DockerInfo {
    // The docker image that is going to be passed to the registry.
    required string image = 1;

    // Network options.
    enum Network {
      HOST = 1;
      BRIDGE = 2;
      NONE = 3;
      USER = 4;
    }

    optional Network network = 2 [default = HOST];

    message PortMapping {
      required uint32 host_port = 1;
      required uint32 container_port = 2;
      // Protocol to expose as (ie: tcp, udp).
      optional string protocol = 3;
    }

    repeated PortMapping port_mappings = 3;

    optional bool privileged = 4 [default = false];

    // Allowing arbitrary parameters to be passed to docker CLI.
    // Note that anything passed to this field is not guaranteed
    // to be supported moving forward, as we might move away from
    // the docker CLI.
    repeated Parameter parameters = 5;

    // With this flag set to true, the docker containerizer will
    // pull the docker image from the registry even if the image
    // is already downloaded on the slave.
    optional bool force_pull_image = 6;

    // The name of volume driver plugin.
    optional string volume_driver = 7 [deprecated = true]; // Since 1.0
  }

  message MesosInfo {
    optional Image image = 1;
  }

  required Type type = 1;
  repeated Volume volumes = 2;
  optional string hostname = 4;

  // Only one of the following *Info messages should be set to match
  // the type.
  optional DockerInfo docker = 3;
  optional MesosInfo mesos = 5;

  // A list of network requests. A framework can request multiple IP addresses
  // for the container.
  repeated NetworkInfo network_infos = 7;

  // Linux specific information for the container.
  optional LinuxInfo linux_info = 8;
}


/**
 * Container related information that is resolved during container
 * setup. The information is sent back to the framework as part of the
 * TaskStatus message.
 */
message ContainerStatus {
  // This field can be reliably used to identify the container IP address.
  repeated NetworkInfo network_infos = 1;

  // Information about Linux control group (cgroup).
  optional CgroupInfo cgroup_info = 2;

  // Information about Executor PID.
  optional uint32 executor_pid = 3;
}


/**
 * Linux control group (cgroup) information.
 */
message CgroupInfo {
  // Configuration of a net_cls cgroup subsystem.
  message NetCls {
    // The 32-bit classid consists of two parts, a 16 bit major handle
    // and a 16-bit minor handle. The major and minor handle are
    // represented using the format 0xAAAABBBB, where 0xAAAA is the
    // 16-bit major handle and 0xBBBB is the 16-bit minor handle.
    optional uint32 classid = 1;
  }

  optional NetCls net_cls = 1;
}


/**
 * Collection of labels. Labels should not contain duplicate key-value
 * pairs.
 */
message Labels {
  repeated Label labels = 1;
}


/**
 * Key, value pair used to store free form user-data.
 */
message Label {
  required string key = 1;
  optional string value = 2;
}


/**
 * Named port used for service discovery.
 */
message Port {
  // Port number on which the framework exposes a service.
  required uint32 number = 1;

  // Name of the service hosted on this port.
  optional string name = 2;

  // Layer 4-7 protocol on which the framework exposes its services.
  optional string protocol = 3;

  // This field restricts discovery within a framework (FRAMEWORK),
  // within a Mesos cluster (CLUSTER), or places no restrictions (EXTERNAL).
  // The visibility setting for a Port overrides the general visibility setting
  // in the DiscoveryInfo.
  optional DiscoveryInfo.Visibility visibility = 4;

  // This can be used to decorate the message with metadata to be
  // interpreted by external applications such as firewalls.
  optional Labels labels = 5;
}


/**
 * Collection of ports.
 */
message Ports {
  repeated Port ports = 1;
}


/**
* Service discovery information.
* The visibility field restricts discovery within a framework (FRAMEWORK),
* within a Mesos cluster (CLUSTER), or places no restrictions (EXTERNAL).
* Each port in the ports field also has an optional visibility field.
* If visibility is specified for a port, it overrides the default service-wide
* DiscoveryInfo.visibility for that port.
* The environment, location, and version fields provide first class support for
* common attributes used to differentiate between similar services. The
* environment may receive values such as PROD/QA/DEV, the location field may
* receive values like EAST-US/WEST-US/EUROPE/AMEA, and the version field may
* receive values like v2.0/v0.9. The exact use of these fields is up to each
* service discovery system.
*/
message DiscoveryInfo {
  enum Visibility {
    FRAMEWORK = 0;
    CLUSTER = 1;
    EXTERNAL = 2;
  }

  required Visibility visibility = 1;
  optional string name = 2;
  optional string environment = 3;
  optional string location = 4;
  optional string version = 5;
  optional Ports ports = 6;
  optional Labels labels = 7;
}


/**
 * Named WeightInfo to indicate resource allocation
 * priority between the different roles.
 */
message WeightInfo {
  required double weight = 1;

  // Related role name.
  optional string role = 2;
}


/**
 * Version information of a component.
 */
message VersionInfo {
  required string version = 1;
  optional string build_date = 2;
  optional double build_time = 3;
  optional string build_user = 4;
  optional string git_sha = 5;
  optional string git_branch = 6;
  optional string git_tag = 7;
}


/**
 * Flag consists of a name and optionally its value.
 */
message Flag {
  required string name = 1;
  optional string value = 2;
}


/**
 * Describes a Role. Roles can be used to specify that certain resources are
 * reserved for the use of one or more frameworks.
 */
message Role {
  required string name = 1;
  required double weight = 2;
  repeated FrameworkID frameworks = 3;
  repeated Resource resources = 4;
}


/**
 * Metric consists of a name and optionally its value.
 */
message Metric {
  required string name = 1;
  optional double value = 2;
}


/**
 * Describes a File.
 */
message FileInfo {
  // Absolute path to the file.
  required string path = 1;

  // Number of hard links.
  optional int32 nlink = 2;

  // Total size in bytes.
  optional uint64 size = 3;

  // Last modification time.
  optional TimeInfo mtime = 4;

  // Represents a file's mode and permission bits. The bits have the same
  // definition on all systems and is portable.
  optional uint32 mode = 5;

  // User ID of owner.
  optional string uid = 6;

  // Group ID of owner.
  optional string gid = 7;
}