Configuring Schedoscope

Internally, Schedoscope uses Typesafe's Akka framework. As a consequence, Schedoscope makes use of the Typesafe Config Library for configuration purposes.

Hence, all configuration properties and their default values are defined in and taken from the file reference.conf on the classpath in the JSON-like HOCON format.

Individual instances of Schedoscope can override these properties. This can happen by either

• putting a file application.conf onto the classpath or
• setting the system property -Dconfig.file pointing to a conf file when launching the JVM.
• for more information regarding Metascope check Metascope Configuration

Reference configuration

Here, you find the commented default configuration of Schedoscope's core as stored in artifact schedoscope-conf's reference.conf file.

At the very minimum, you should check and set if necessary:

• the environment name (schedoscope.app.environment);
• the earliest day to consider (schedoscope.scheduler.earliestDay);
• the Resource Manager, Namenode, and Metastore hosts (schedoscope.hadoop.*, schedoscope.metastore.*);
• the Kerberos Metastore principal (schedoscope.kerberos.principal);
• the Hive JDBC URL (schedoscope.metastore.jdbcUrl )
• the number of concurrent driver actors for each transformation type (schedoscope.transformations.*.concurrency);
• the future-driver-dispatcher threadpool configuration depending on your concurrency settings.

For an example of how to override these settings, you can also take a look at the [tutorial](Open Street Map Tutorial).

reference.conf (schedoscope-conf)

  #
  # Schedoscope configuration properties and default values
  #

  schedoscope {

    #
    # Basic settings
    #

    app {

      #
      # The environment name of your Schedoscope instance. This allows you to run
      # multiple instances of schedoscope, such as for development, testing or
      # production but also for different applications / datawarehouses.
      #
      # The environment impacts the default database and partition path naming
      # schemes.
      #
      # Common environment names are: dev, int, prod
      #

      environment = "dev"

      #
      # View augmentor (a class extending
      # org.schedoscope.dsl.views.ViewUrlParser.ParsedViewAugmentor)
      # to use for dynamic view instantiation via the view pattern URL scheme.
      # This can be used to modify instantiated views in application-specific
      # ways outside of view pattern URLs. Should rarely need change.
      #

      parsedViewAugmentorClass = "org.schedoscope.dsl.views.NoAugmentation"
    }

    #
    # Settings with regard to scheduling logic
    #

    scheduler {

      #
      # The earliest date to consider for MonthlyParameterization or
      # DailyParameterization view traits. Should be the earliest day you have
      # data for your application to avoid any needless no-data overhead.
      #

      earliestDay = "2013-12-01"

      #
      # The latest date to consider for MonthlyParameterization or
      # DailyParameterization view traits. Should usually be "now", but can be
      # reduced to a fixed date, for example in testing environments.
      #

      latestDay = "now"


      #
      # Various timeouts to obey
      #

      timeouts {

	#
	# Timeout waiting for metastore operations
	#

	schema = 30 minutes


	#
	# Timeout waiting for scheduling command to return
	#

	schedulingCommand = 30 minutes
      }

      #
      # Registered listeners for view scheduling state changes
      #

      viewSchedulingListeners = ["org.schedoscope.scheduler.listeners.ViewSchedulingMonitor"]

    }

    #
    # Configuration properties with regard Schedoscope's REST web service API.
    # Used for both Schedoscope server and the Schedoscope client tool.
    #

    webservice {


      #
      # Host where Schedoscope instance is running
      #

      host = "localhost"

      #
      # Port on which the Schedoscope web service listens for connections
      #

      port = 20698

      #
      # Directory for storing static web service resources
      #

      resourceDirectory = "www"

      #
      # Number of webservice actors to answer requests.
      #

      concurrency = 5
    }


    #
    # Kerberos-related settings
    #

    kerberos {

      #
      # The kerberos principal required for access to the Metastore
      # E.g.: hive/somemachine@SOMEDOMAIN.COM
      #
      principal = ""
    }

    #
    # Settings related to core Hadoop
    #

    hadoop {

      #
      # Address and port of the YARN resource manager. Note that any locally configured Hadoop environment
      # takes precedence.
      #

      resourceManager = "localhost:8032"

      #
      # Address and port of the HDFS namenode. Note that any locally configured Hadoop environment
      # takes precedence.
      #

      nameNode = "localhost:8020"

      #
      # HDFS URI
      #

      hdfs = "hdfs://localhost:8020"

      #
      # Schedoscope stores view data by default below this root folder.
      #

      viewDataHdfsRoot = "/hdp"
    }


    #
    # Settings related to dev environment
    #

    development {

      #
      # Enables the development feature
      #
      enabled = false

      #
      # Adress of the production namenode
      #
      prodNameNode = "locahost:8020"

      #
      # The environment under prod
      #
      prodEnv = "prod"

      #
      # Hdfs root on prod
      #
      prodViewDataHdfsRoot = "/hdp"

      #
      # Enables the usage of ssh to run the distcp on a remote machine
      #
      sshEnabled = false

      #
      # Ssh to this machine. Only needed if ssh is enabled
      #
      sshTarget = ""

      #
      # Views under development i.e. views which dependencies will be stubbed
      # by loading data from the prod environment.
      #
      # You can define one view or a list of views.
      #
      # e.g.: [ schedoscope.example.osm.datamart/ShopProfiles ]
      #
      viewUrls = [""]

    }

    #
    # Settings related to the Hive metastore
    #

    metastore {

      #
      # Thrift URI to metastore service
      #
      metastoreUri = "thrift://localhost:9083"

      #
      # Hive JDBC URL to connect to Hive Server 2
      #

      jdbcUrl = "jdbc:hive2://localhost:10000/default"

      #
      # Path to hive config files on your cluster
      #

      hiveConfDir = "/etc/hive/conf/"

      #
      # Number of parallel connections to use for schema- and partition creation
      # via the Hive metastore
      #

      concurrency = 5

      #
      # Back-off slot time (millis) for Actors dealing with metastore
      #
      actor-backoff-slot-time = 20000

      #
      # Back-off minimum constant delay (millis) for Actors dealing with metastore
      #
      actor-backoff-minimum-delay = 20000

      #
      # Number of partitions to put into a batch for each connection when
      # creating partitions
      #

      writeBatchSize = 2500

      #
      # Number of partitions to read in batch with each request on each
      # connection
      #

      readBatchSize = 10000
    }

    #
    # Settings concerning automatic logic change detection
    #

    versioning {

      #
      # Compute transformation version checksums and detect need for
      # rematerialization based on those checksums. Should be kept as is.
      #

      transformations = true
    }

    #
    # Number of retries for transformations before view state is changed to
    # failed.
    #

    action.retry = 5

    #
    # Setting concerning the usage of external dependencies
    #

    external-dependencies {

      #
      # This setting allows you to use external dependencies and operate several schedoscope instances in conjunction.
      #

      enabled = false

      #
      # A list of prefixes of packages with internal views. Every package not starting with a string in this list
      # will be treated as external and can not be referenced from the client or used as dependency if not flagged as
      # exernal
      #

      home = ["${env}.test.datahub"]

      #
      # Toggles checks whether internal views are used as external views and vice versa
      #

      checks = true
    }

    #
    # Default settings for exportTo() exporters.
    #

    export {

      #
      # Globally disable exports
      #

      disableAll = false

      #
      # Export settings for all export types
      #

      salt = "vD75MqvaasIlCf7H"

      #
      # BigQuery exporter settings.
      #

      bigQuery {

	#
	# GCP project ID under which the exported BigQuery dataset will be created
	#

	projectId = ""

	#
	# GCP key in JSON format to use for authentication
	#

	gcpKey = ""

	#
	# Number of reducers to use for parallel writing to BigQuery.
	#

	numberOfReducers = 10


	#
	# Number of records to batch before flushing data to GCP Cloud Storage.
	#

	flushInterval = 10000

	#
	# GCP data storage location of exported data within BigQuery.
	#

	dataLocation = "EU"

	#
	# GCP Cloud Storage bucket for temporary storage to use for exporting to BigQuery.
	#

	exportStorageBucket = "schedoscope-bigquery-export"

	#
	# GCP Cloud Storage bucket folder prefix to apply to blobs when exporting to BigQuery
	#

	exportStorageBucketFolderPrefix = ""

	#
	# GCP Cloud Storage bucket region to use for exporting to BigQuery
	#

	exportStorageBucketRegion = "europe-west3"

	#
	# Host of proxy to use for GCP API access
	#

	proxyHost = ""

	#
	# Port of proxy to use for GCP API access
	#

	proxyPort = ""


      }

      #
      # JDBC exporter settings.
      #

      jdbc {

	#
	# Number of reducers to use for parallel writing to JDBC.
	#

	numberOfReducers = 10

	#
	# Batch size for insert statements to use
	#

	insertBatchSize = 10000

	#
	# Storage engine to use (in case of MySQL JDBC targets)
	#

	storageEngine = "InnoDB"

      }

      #
      # Redis exporter.
      #

      redis {

	#
	# Number of reducers to use for parallel writing to Redis.
	#

	numberOfReducers = 10

	#
	# Usage of pipeline mode for writing to Redis.
	#

	usePipelineMode = false

	#
	# Batch size for Redis pipeline mode
	#

	insertBatchSize = 10000

      }

      #
      # Kafka exporter
      #

      kafka {

	#
	# Number of reducer to use for parallel writing to Kafka.
	#

	numberOfReducers = 10

      }

      #
      # Ftp exporter
      #

      ftp {

	#
	# Number of reducers to use for parallel exporting to a (S)Ftp Server.
	#

	numberOfReducers = 2

      }
    }

    #
    # Metascope related settings
    #

    metascope {

      #
      # Metascope webservice port
      #

      port = 8080

      #
      # The Schedoscope instance(s) webservice to retrieve the metadata from
      #

      schedoscope-instances = [
	{
	  id = "schedoscope-1"
	  host = "localhost"
	  port = 20698
	}
      ]

      #
      # Configure authentication method
      #

      auth {

	#
	# Possible methods = ["simple", "ldap"]
	# simple: custom user management
	# ldap: external ldap directory is used for authentication
	#

	authentication = "simple"

	#
	# ldap settings
	#

	ldap {

	  #
	  # ldap server url
	  #

	  url = ""

	  #
	  #	ldap manager dn
	  #

	  managerDn = ""

	  #
	  # ldap manager password
	  #

	  managerPassword = ""

	  #
	  # ldap user dn pattern
	  #

	  userDnPattern = ""

	  #
	  # ldap group search base
	  #

	  groupSearchBase = ""

	  #
	  # Allowed groups to access Metascope
	  #

	  allowedGroups = ""

	  #
	  #	Admin group(s): User will have admin rights
	  #

	  adminGroups = ""

	}

      }

      #
      # Database which is used as the metadata repository.
      # Default: embedded Apache Derby instance, which creates a local repository in the metascope directory ('.../metascope/repository')
      #

      repository {

	#
	# JDBC URL to database server
	#

	url = "jdbc:h2:{metascope.dir}/repository;MODE=MySQL;CREATE=TRUE"

	#
	# database user
	#

	user = "sa"

	#
	# users password
	#

	password = "sa"

	#
	# Dialect to use to talk to database. For further information, visit:
	# http://docs.jboss.org/hibernate/orm/3.5/javadocs/org/hibernate/dialect/package-summary.html
	#

	dialect = "org.hibernate.dialect.H2Dialect"
      }

      #
      # Configure Hive Metastore access
      #

      metastore {

	#
	# Different ways to access metastore (only "thrift" and "jdbc" are allowed
	# "jdbc" mode only works with MySQL. When using "jdbc" mode, the metastore is accessed directly via JDBC
	#

	method = "thrift"

	jdbc {

	  #
	  # JDBC url to Hive Metastore (MySQL only)
	  #

	  url = ""

	  #
	  # database user to access metastore
	  #

	  user = ""

	  #
	  # database user password
	  #

	  password = ""

	}

      }

      #
      # Settigns relataed to the Apache Solr instance
      #

      solr {

	#
	# URL to Solr instance.
	# Default: embedded Apache Solr instance, which creates a local index in the metascope directory ('.../metascope/solr')
	#

	url = "{metascope.dir}/solr"
      }

      #
      # General settings concerning logging (via SLF4J)
      #

      logging {

	#
	# Path to logfile
	#

	logfile = "{metascope.dir}/logs/metascope-service.log"

	#
	# Log level granularity
	#

	loglevel = "DEBUG"
      }

    }

    #
    # Driver settings for the different transformation types.
    #

    transformations = {

      noop: {
	#
	# Class implementing the NoOp driver
	#
	driverClassName = "org.schedoscope.scheduler.driver.NoOpDriver"

	#
	# ignored
	#
	location = "/tmp/schedoscope/noop/"

	#
	# concurrency for NoOp transformations
	#
	concurrency = 5

	#
	# Back-off slot time (millis) for Driver Actor responsible
	#
	driver-actor-backoff-slot-time = 200

	#
	# Back-off minimum constant delay (millis) for Driver Actor responsible
	#
	driver-actor-backoff-minimum-delay = 100

	#
	# Timeout for NoOp transformations.
	#

	timeout = 1 day

	#
	# The handlers being notified after each driver run has
	# finished (succeeded or failed). These must implement the
	# trait org.schedoscope.scheduler.driver.DriverRunCompletionHandler
	#

	driverRunCompletionHandlers = ["org.schedoscope.scheduler.driver.DoNothingCompletionHandler"]

      }

      #
      # Settings for Hive transformations
      #

      hive: {

	#
	# Class implementing the Hive driver
	#

	driverClassName = "org.schedoscope.scheduler.driver.HiveDriver"

	#
	# Location where to put Hive UDFS in HDFS
	#

	location = "/tmp/schedoscope/hive/"

	#
	# Comma-separated list of directories where additional UDF library
	# jars can be found that are to be put onto HDFS when launching
	# Schedoscope.
	#

	libDirectory = ""

	#
	# Ignored
	#

	url = ""

	#
	# Do not change. Hive UDF jars should not be unpacked in HDFS.
	#

	unpack = false

	#
	# Number of parallel Driver actors to use for executing Hive
	# transformations
	#

	concurrency = 10

	#
	# Back-off slot time (millis) for Driver Actor responsible
	#
	driver-actor-backoff-slot-time = 5000

	#
	# Back-off minimum constant delay (millis) for Driver Actor responsible
	#
	driver-actor-backoff-minimum-delay = 3000

	#
	# Timeout for Hive transformations.
	#

	timeout = 1 day

	#
	# The handlers being notified after each driver run has
	# finished (succeeded or failed). These must implement the
	# trait org.schedoscope.scheduler.driver.DriverRunCompletionHandler
	#

	driverRunCompletionHandlers = ["org.schedoscope.scheduler.driver.DoNothingCompletionHandler"]

      },

      #
      # Settings for MapReduce transformations
      #

      mapreduce: {

	#
	# Class implementing the Mapreduce driver
	#

	driverClassName = "org.schedoscope.scheduler.driver.MapreduceDriver"

	#
	# Ignored.
	#

	location = "/tmp/schedoscope/mapreduce/"


	#
	# Ignored.
	#

	libDirectory = ""

	#
	# Ignored.
	#

	url = ""

	#
	# Ignored
	#

	unpack = false

	#
	# Number of parallel Driver actors to use for executing MapReduce
	# transformations
	#

	concurrency = 10

	#
	# Back-off slot time (millis) for Driver Actor responsible
	#
	driver-actor-backoff-slot-time = 5000

	#
	# Back-off minimum constant delay (millis) for Driver Actor responsible
	#
	driver-actor-backoff-minimum-delay = 3000

	#
	# Timeout for MapReduce transformations
	#

	timeout = 1 day

	#
	# The handlers being notified after each driver run has
	# finished (succeeded or failed). These must implement the
	# trait org.schedoscope.scheduler.driver.DriverRunCompletionHandler
	#

	driverRunCompletionHandlers = ["org.schedoscope.scheduler.driver.DoNothingCompletionHandler"]

      },

      #
      # File system driver settings
      #

      filesystem: {

	#
	# Class implementing the FileSystem driver
	#

	driverClassName = "org.schedoscope.scheduler.driver.FilesystemDriver"

	#
	# Number of parallel Driver actors to use for executing file system
	# transformations
	#

	concurrency = 1

	#
	# Back-off slot time (millis) for Driver Actor responsible
	#
	driver-actor-backoff-slot-time = 200

	#
	# Back-off minimum constant delay (millis) for Driver Actor responsible
	#
	driver-actor-backoff-minimum-delay = 100

	#
	# Ignored.
	#

	location = "/tmp/schedoscope/filesystem/"

	#
	# Ignored
	#

	libDirectory = ""

	#
	# Ignored
	#

	url = ""

	#
	# Ignored
	#

	unpack = false

	#
	# Timeout for file system transformations
	#

	timeout = 1 hour

	#
	# The handlers being notified after each driver run has
	# finished (succeeded or failed). These must implement the
	# trait org.schedoscope.scheduler.driver.DriverRunCompletionHandler
	#

	driverRunCompletionHandlers = ["org.schedoscope.scheduler.driver.DoNothingCompletionHandler"]

      },

      #
      # Sequence driver settings
      #

      seq: {

	#
	# Class implementing the Seq driver
	#

	driverClassName = "org.schedoscope.scheduler.driver.SeqDriver"

	#
	# Number of parallel Driver actors to use for executing seq
	# transformations
	#

	concurrency = 10

	#
	# Back-off slot time (millis) for Driver Actor responsible
	#
	driver-actor-backoff-slot-time = 200

	#
	# Back-off minimum constant delay (millis) for Driver Actor responsible
	#
	driver-actor-backoff-minimum-delay = 100

	#
	# Ignored.
	#

	location = "/tmp/schedoscope/seq/"

	#
	# Ignored
	#

	libDirectory = ""

	#
	# Ignored
	#

	url = ""

	#
	# Ignored
	#

	unpack = false

	#
	# Timeout for seq transformations
	#

	timeout = 1 day

	#
	# The handlers being notified after each driver run has
	# finished (succeeded or failed). These must implement the
	# trait org.schedoscope.scheduler.driver.DriverRunCompletionHandler
	#

	driverRunCompletionHandlers = ["org.schedoscope.scheduler.driver.DoNothingCompletionHandler"]

      }

      #
      # Shell driver settings
      #

      shell: {

	#
	# Class implementing the Shell driver
	#

	driverClassName = "org.schedoscope.scheduler.driver.ShellDriver"

	#
	# Number of parallel Shell Driver actors to use
	#

	concurrency = 1

	#
	# Back-off slot time (millis) for Driver Actor responsible
	#
	driver-actor-backoff-slot-time = 200

	#
	# Back-off minimum constant delay (millis) for Driver Actor responsible
	#
	driver-actor-backoff-minimum-delay = 0

	#
	# Ignored
	#

	location = "/tmp/schedoscope/shell/"

	#
	# Ignored
	#

	libDirectory = ""

	#
	# Ignored
	#

	url = ""

	#
	# Ignored
	#

	unpack = false

	#
	# Timeout for Shell transformations
	#

	timeout = 1 day

	#
	# The handlers being notified after each driver run has
	# finished (succeeded or failed). These must implement the
	# trait org.schedoscope.scheduler.driver.DriverRunCompletionHandler
	#

	driverRunCompletionHandlers = ["org.schedoscope.scheduler.driver.DoNothingCompletionHandler"]
      }
    }

    test {
      disableLineageValidation = true
    }

  }

  #
  # This section covers general Akka settings, in particular concerning the
  # various dispatchers / threadpools.
  #

  akka {

    #
    # General settings concerning logging (via SLF4J). Logback is given as a
    # default dependency to Schedoscope so SLF4J via Logback is the default
    # logging system.
    #

    log-config-on-start = off
    log-dead-letters = off
    loglevel = "INFO"
    loggers = ["akka.event.slf4j.Slf4jLogger"]

    #
    # Dispatcher settings
    #

    actor {

      guardian-supervisor-strategy = "org.schedoscope.TerminatingStoppingStrategy"

      #
      # Given the variety of component-specific dispatchers used by Schedoscope
      # we limit the Akka default dispatcher to 4 threads.
      #

      default-dispatcher {
	executor = "thread-pool-executor"
	type = Dispatcher

	thread-pool-executor {
	  core-pool-size-min = 4
	  core-pool-size-factor = 1.0
	  core-pool-size-max = 4
	  task-queue-size = -1
	}

	throughput = 5
      }


      #
      # Configuration of threads to use for unspecific future calls and
      # computations.
      # Again limited.
      #

      future-call-dispatcher {
	type = Dispatcher
	executor = "thread-pool-executor"

	thread-pool-executor {
	  core-pool-size-min = 8
	  core-pool-size-factor = 2.0
	  core-pool-size-max = 8
	  task-queue-size = -1
	}

	throughput = 5
      }


      #
      # The supervisor / message router for the Metastore-related actors is
      # assigned to one pinned dispatcher / thread. Should not need to be
      # changed.
      #

      schema-manager-dispatcher {
	executor = "thread-pool-executor"
	type = PinnedDispatcher
      }

      #
      # Threadpool / dispatcher configuration for the actor communicating
      # with the Metastore for creating / reading tables and partitions. Should
      # be aligned with the property schedoscope.metastore.concurrency
      #

      partition-creator-dispatcher {
	executor = "thread-pool-executor"
	type = Dispatcher
	mailbox-type = "org.schedoscope.scheduler.actors.ManagedActorPriorityMailbox"

	thread-pool-executor {
	  core-pool-size-min = 5
	  core-pool-size-factor = 1.0
	  core-pool-size-max = 5
	  task-queue-size = -1
	}

	throughput = 5
      }


      #
      # The actor logging transformation metadata (version, timestamp) to the
      # Metastore has one dedicated pinned dispatcher / thread.
      #

      metadata-logger-dispatcher {
	executor = "thread-pool-executor"
	type = PinnedDispatcher
	mailbox-type = "org.schedoscope.scheduler.actors.ManagedActorPriorityMailbox"
      }

      #
      # The actor managing, assigning, and collecting status information about
      # transformations to execute has one dedicated pinned dispatcher / thread
      # for reasons of responsiveness.
      #

      transformation-manager-dispatcher {
	executor = "thread-pool-executor"
	type = PinnedDispatcher
      }

      #
      # The actor managing and instantiating view actors for scheduling state
      # management has one dedicated pinned dispatcher / thread for reasons of
      # responsiveness.
      #

      view-manager-dispatcher {
	executor = "thread-pool-executor"
	type = PinnedDispatcher
      }

      #
      # The threadpool / dispatcher used by the table actors for communicating
      # with each other for view materialization and scheduling.
      #

      views-dispatcher {
	executor = "fork-join-executor"
	type = Dispatcher

	fork-join-executor {
	  parallelism-min = 8
	  parallelism-factor = 2.0
	  parallelism-max = 8
	  task-queue-size = -1
	}

	throughput = 5
      }

      #
      # The Router actor managing and instantiating driver actors
      # has one dedicated pinned dispatcher / thread for reasons of
      # responsiveness.
      #

      driver-router-dispatcher {
	executor = "thread-pool-executor"
	type = PinnedDispatcher
      }

      #
      # The threadpool / dispatcher available for the transformation drivers.
      # As drivers execute transformations asynchronously - either by
      # communicating with servers like Oozie or the Resource Manager or by
      # futures, this pool should not need to grow much when increasing
      # concurrency settings of the various transformations.
      #

      driver-dispatcher {
	executor = "thread-pool-executor"
	type = Dispatcher
	mailbox-type = "org.schedoscope.scheduler.actors.ManagedActorPriorityMailbox"

	thread-pool-executor {
	  core-pool-size-min = 8
	  core-pool-size-factor = 2.0
	  core-pool-size-max = 8
	  task-queue-size = -1
	}

	throughput = 5
      }

      #
      # Due to API limitations, we can currently only execute file system, Pig,
      # and Hive transformations asynchronously by employing futures.
      #
      # The following threadpool / dispatcher feeds those futures. Hence, it
      # should grow proportionally with the concurrency settings of the
      # mentioned transformation types.
      #

      future-driver-dispatcher {
	type = Dispatcher
	executor = "thread-pool-executor"

	thread-pool-executor {
	  core-pool-size-min = 16
	  core-pool-size-factor = 4.0
	  core-pool-size-max = 16
	  task-queue-size = -1
	}

	throughput = 5
      }

      #
      # The Supervising strategy for view scheduling Listener Actors
      # maxRetries - maximum number of restart attempts for OneForOneStrategy
      #
      view-scheduling-listener-actor {
	maxRetries = -1
      }

    }
  }

  #
  # The following are a list of "suggestions" for spray settings to use for the
  # Schedoscope web service and client. They are not automatically pulled
  # but have to be explicitly set in your application.conf / schedoscope.conf.
  #

  spray.can {
    client {
      parsing {
	max-content-length = 100m
      }

      # The time after which an idle connection will be automatically closed.
      # Set to `infinite` to completely disable idle timeouts.
      idle-timeout = 1200s

      #  The max time period that a client connection will be waiting for a response
      #  before triggering a request timeout. The timer for this logic is not started
      #  until the connection is actually in a state to receive the response, which
      #  may be quite some time after the request has been received from the
      #  application!
      #  There are two main reasons to delay the start of the request timeout timer:
      #    1. On the host-level API with pipelining disabled:
      #  If the request cannot be sent immediately because all connections are
      #  currently busy with earlier requests it has to be queued until a
      #  connection becomes available.
      #  2. With pipelining enabled:
      #    The request timeout timer starts only once the response for the
      #  preceding request on the connection has arrived.
      #  Set to `infinite` to completely disable request timeouts.
      request-timeout = 900s

      # The time period within which the TCP connecting process must be completed.
      # Set to `infinite` to disable.
      connecting-timeout = 10s

    }
    server {
      # If a request hasn't been responded to after the time period set here
      # a `spray.http.Timedout` message will be sent to the timeout handler.
      # Set to `infinite` to completely disable request timeouts.
      request-timeout = 900s

      # After a `Timedout` message has been sent to the timeout handler and the
      # request still hasn't been completed after the time period set here
      # the server will complete the request itself with an error response.
      # Set to `infinite` to disable timeout timeouts.
      timeout-timeout = 6s

      # The time after which an idle connection will be automatically closed.
      # Set to `infinite` to completely disable idle connection timeouts.
      idle-timeout = 1200s

      # Enables/disables the returning of more detailed error messages to
      # the client in the error response.
      # Should be disabled for browser-facing APIs due to the risk of XSS attacks
      # and (probably) enabled for internal or non-browser APIs.
      # Note that spray will always produce log messages containing the full
      # error details.
      verbose-error-messages = off

      # Enables/disables the logging of the full (potentially multiple line)
      # error message to the server logs.
      # If disabled only a single line will be logged.
      verbose-error-logging = off

      # Enables/disables support for statistics collection and querying.
      # Even though stats keeping overhead is small,
      # for maximum performance switch off when not needed.
      stats-support = on

      # Enables/disables automatic back-pressure handling by write buffering and
      # receive throttling
      automatic-back-pressure-handling = on
      back-pressure {
	# The reciprocal rate of requested Acks per NoAcks. E.g. the default value
	# '10' means that every 10th write request is acknowledged. This affects the
	# number of writes each connection has to buffer even in absence of back-pressure.
	noack-rate = 10
	# The lower limit the write queue size has to shrink to before reads are resumed.
	# Use 'infinite' to disable the low-watermark so that reading is resumed instantly
	# after the next successful write.
	reading-low-watermark = infinite
      }
    }
  }