3.7.6 a maintenance release that primarily includes bug fixes.
Upgrades and Compatibility
This release includes one potential incompatibility with earlier
3.7.x releases. Only systems
that use a high number (hundreds) of queue priorities are affected.
See 3.7.0 release notes upgrade and compatibility notes if upgrading from an earlier release.
See the Upgrading guide for general documentation on upgrades.
See RabbitMQ change log for release notes of other releases.
Max priority cap for queues is now enforced and set to 255. Applications that rely on a higher number of priorities will break.
Such applications must be updated to use no more than 255 priorities.
GitHub issue: rabbitmq-server#1590
Bulk deletion of exclusive queues when their owning connection closes has been optimized.
On Erlang 20.2.3 or later, more optimal memory allocators will be used. On some workloads that leads
to significant reduction in node's RAM consumption. The change is workload-specific, however, on the workloads
tested that do introduce a negative effect (more RAM used) the difference was very small (about 1%).
It is possible to go back to the previous settings,
+MBas aoffcbf +MHas aoffcbf +MBlmbcs 5120 +MHlmbcs 5120 +MMmcs 10,
by specifying them in the
On Erlang/OTP releases older than 20.2.3 there will be no default changes.
GitHub issue: rabbitmq/rabbitmq-server#1604
Logging to Syslog without 3rd party plugins.
GitHub issue: rabbitmq-server#1600
connection.closedevents now include client-provided connection name and more client details.
That information is in turn available via the rabbitmq_event_exchange.
GitHub issue: rabbitmq-server#1596
await_online_nodes <count> [--timeout <timeout>]is a new command that will wait until the target
node's cluster has at least
<count>nodes. The command supports a timeout. It is meant to simplify
cluster provisioning automation and can be used as the very first check performed after starting a node.
Note that the command requires at least one node to be running and reachable.
GitHub issue: rabbitmq-cli#235
rabbitmq-diagnostics ping [--node <target> --timeout <timeout>]is a new minimalistic health check command.
It only verifies if the target node is up, registered with epmd and accepts/authenticates CLI tool connections.
It fills the gap between
rabbitmq-diagnostics node_health_check, which performs a lot of checks, take a lot more time
and is fairly opinionated in what it does.
pingis more suitable in environments where
failing aliveness checks will result in node restarts, namely Kubernetes.
GitHub issue: rabbitmq-cli#253
Management and Management Agent Plugins
Queue metrics could be missing after a node restart or plugin [re-]enablement.
GitHub issue: rabbitmq/rabbitmq-management-agent#58
rabbitmqadminnow correctly percent-encodes spaces in connection names.
GitHub issue: rabbitmq/rabbitmq-management#576
When creating a new vhost, user permissions will only be granted to the users that exist in the
internal database. If an external authentication/authorization service is used, the step is skipped.
GitHub issue: rabbitmq-management#531
Virtual host imported from definitions or created via HTTP API will explicitly await for the vhost to initialise
on all cluster nodes.
GitHub issue: rabbitmq/rabbitmq-management#575
LDAP connection pool now correctly purges connections in more failure scenarios (e.g. LDAP server restart)
GitHub issue: rabbitmq-auth-backend-ldap#92
HTTP Authentication/Authorization Plugin
201 Createdresponses are now accepted from HTTP services that back the plugin.
GitHub issue: rabbitmq-auth-backend-http#68
Source code archives
Warning: The source code archive provided by GitHub only contains the source of the broker,
not the plugins or the client libraries. Please download the archive named