Actuator Endpoints

最后更新于2年前

Actuator Endpoints

For monitoring, AxonServer includes Spring Boot Actuator endpoints which are available under the /actuator URL path of the AxonServer.‌ For Axon Server SE, the URL for the Axon Server SE will be the single running node, while for Axon Server EE the URL should be pointing to any node serving the _admin context within an Axon Server EE cluster.

The /actuatorURL path returns a list of all available actuators. A list of the endpoints is given below.

/actuator/health endpoint is used to check the health of Axon Server itself, and (in the case of Axon Server EE) the health of the cluster. The HTTP status return code is not 200 when cluster health is anything other than "UP".
/actuator/info endpoint informs you about some basic attributes of an AxonServer (name, description, version). This is more useful for liveness/readiness probes.
/actuator/metrics endpoint publishes information about OS, JVM as well as
/actuator/loggers endpoint exposes detailed view of the loggers configuration
/actuator/prometheus endpoint exposes metrics data in a format that can be scraped by a
/actuator/env endpoint exposes the current environment properties

Health endpoint

The /actuator/health endpoint provides information on various components of Axon Server:

cluster (EE only), shows the status of the connection between the current Axon Server node and other nodes in the cluster. This section also shows information on the flow control between Axon Server nodes. If the value for commands.waiting or queries.waiting is not zero for a longer period of time, it means that the connected Axon Server node cannot process the commands or queries fast enough. If there are waiting commands or queries, or the connection with one of the other nodes is down, this component will show status WARN. If the connections to all other nodes is down, the status of the cluster component is DOWN.
command, shows the status of command handling applications connected to the current Axon Server node. For each connected application it shows the number of commands waiting to be sent to the command handler. If this value is higher than zero, it means that commands are being sent faster than the command handler can handle them.
db, shows the status of the control DB
diskSpace, shows the diskspace in use and available for each storage location (this includes the locations where events and snapshots are stored, and the location of the replication logs). if the available space in any of these locations is below the threshold, the health for the diskSpace component will be WARN.
license (EE only, since 4.5). shows the status of the license. If the license is in the grace period, this component gets status WARN, if license is expired the status becomes DOWM.
localEventStore, shows the status of the local replica for each context available on the current node.
ping, always shows UP
query, shows the status of query handling applications connected to the current Axon Server node. For each connected application it shows the number of queries waiting to be sent to the query handler. If this value is higher than zero, it means that queries are being sent faster than the query handler can handle them.
raft, shows the status of the replication process per replication group. It shows the leader of the replication group. For the leader it shows the status of the followers, checking if the follower has recently confirmed messages and if the follower is not too far behind. If there is a problem with the replication on one of the replication groups the health for this component is WARN. If there are issues for all replication groups the status is DOWN.

The overall health status is derived from the status of the components. If one of the components is WARN, the overall status is WARN. If one of the components is DOWN, it is DOWN.

The default setting for the health endpoint in Axon Server is to show the details, even when the user is not authenticated. To hide the details for non-authenticated users add this property to the axonserver.properties file:

management.endpoint.health.show-details=when-authorized

Application level metrics‌

Specific AxonServer metrics are available under /actuator/metrics endpoint:‌

/actuator/metrics/axon.events.count current number of events stored in a node
/actuator/metrics/axon.commands.count current number of commands stored in a node
/actuator/metrics/axon.queries.count current number of queries stored in a node
/actuator/metrics/axon.snapshots.count current number of snapshots stored in a node
/actuator/metrics/axon.commands.active current number of active commands in a node

上一页Monitoring 下一页gRPC Metrics

最后更新于2年前

The /actuatorURL path returns a list of all available actuators. A list of the endpoints is given below.

/actuator/health endpoint is used to check the health of Axon Server itself, and (in the case of Axon Server EE) the health of the cluster. The HTTP status return code is not 200 when cluster health is anything other than "UP".
/actuator/info endpoint informs you about some basic attributes of an AxonServer (name, description, version). This is more useful for liveness/readiness probes.
/actuator/metrics endpoint publishes information about OS, JVM as well as
/actuator/loggers endpoint exposes detailed view of the loggers configuration
/actuator/prometheus endpoint exposes metrics data in a format that can be scraped by a
/actuator/env endpoint exposes the current environment properties

Health endpoint

The /actuator/health endpoint provides information on various components of Axon Server:

cluster (EE only), shows the status of the connection between the current Axon Server node and other nodes in the cluster. This section also shows information on the flow control between Axon Server nodes. If the value for commands.waiting or queries.waiting is not zero for a longer period of time, it means that the connected Axon Server node cannot process the commands or queries fast enough. If there are waiting commands or queries, or the connection with one of the other nodes is down, this component will show status WARN. If the connections to all other nodes is down, the status of the cluster component is DOWN.
command, shows the status of command handling applications connected to the current Axon Server node. For each connected application it shows the number of commands waiting to be sent to the command handler. If this value is higher than zero, it means that commands are being sent faster than the command handler can handle them.
db, shows the status of the control DB
diskSpace, shows the diskspace in use and available for each storage location (this includes the locations where events and snapshots are stored, and the location of the replication logs). if the available space in any of these locations is below the threshold, the health for the diskSpace component will be WARN.
license (EE only, since 4.5). shows the status of the license. If the license is in the grace period, this component gets status WARN, if license is expired the status becomes DOWM.
localEventStore, shows the status of the local replica for each context available on the current node.
ping, always shows UP
query, shows the status of query handling applications connected to the current Axon Server node. For each connected application it shows the number of queries waiting to be sent to the query handler. If this value is higher than zero, it means that queries are being sent faster than the query handler can handle them.
raft, shows the status of the replication process per replication group. It shows the leader of the replication group. For the leader it shows the status of the followers, checking if the follower has recently confirmed messages and if the follower is not too far behind. If there is a problem with the replication on one of the replication groups the health for this component is WARN. If there are issues for all replication groups the status is DOWN.

The overall health status is derived from the status of the components. If one of the components is WARN, the overall status is WARN. If one of the components is DOWN, it is DOWN.

management.endpoint.health.show-details=when-authorized

Application level metrics‌

Specific AxonServer metrics are available under /actuator/metrics endpoint:‌

/actuator/metrics/axon.events.count current number of events stored in a node
/actuator/metrics/axon.commands.count current number of commands stored in a node
/actuator/metrics/axon.queries.count current number of queries stored in a node
/actuator/metrics/axon.snapshots.count current number of snapshots stored in a node
/actuator/metrics/axon.commands.active current number of active commands in a node