Cluster management beans
As of Jira 7.3, some aspects of each node in a Jira Data Center environment can be monitored via Java Management Extensions (JMX). This can be useful for integrating with third-party monitoring tools, such as Nagios XI, for tracking and visualizing the progress of a zero downtime upgrade.
Note that these interfaces are considered Experimental API, and are subject to modification or removal in later releases.
Before you begin
Ensure that each node in your deployment has a JMX Agent running, and is accessible from the host you intend to monitor from. See Monitoring and Management Using JMX Technology regarding JMX Agents.
Remember to take the security considerations into account when exposing these interfaces.
What you need to do
Ensure that the Cluster Monitoring Dark Feature is enabled. This can be verified by logging in as an administrator, and going to the
jira.zdu.jmx-monitoring should be listed under either System Property Dark Features or Site Wide Dark Features.
If cluster monitoring is not enabled, it must be enabled separately for each node in the deployment. This can be done by passing a JVM system property at startup:
This will require a restart. See Setting properties and options on startup.
Once Jira has started, an MBean will exist at the path
com.atlassian.jira.cluster.monitoring.ClusterNodeState. This can be viewed with a JMX Monitoring Tool such as
ClusterNodeState exposes the following attributes:
NodeVersion: The Jira Platform application version that is currently installed and running on the node. Note that this value refers to the Jira binary version, and not the database schema version. Our current Zero Downtime Upgrade process dictates that the database is upgraded after the binaries have been updated on each node.
ClusterUpgradeState: The state of the node as it pertains to a Zero Downtime Upgrade.
These can then be queried and visualised in a monitoring tool of your choice.
Was this helpful?Yes Provide feedback about this article