Prometheus metrics

ThingsBoard Edge exposes internal metrics via Spring Boot Actuator using the Prometheus format. These metrics can be scraped by a Prometheus instance and visualized in Grafana or any compatible monitoring tool.

Enable Prometheus metrics

Add the following environment variables to your Edge configuration:

export METRICS_ENABLED=true
export METRICS_ENDPOINTS_EXPOSE=prometheus

For Docker deployments, add these to the environment block in docker-compose.yml:

      METRICS_ENABLED: "true"
      METRICS_ENDPOINTS_EXPOSE: "prometheus"

Metrics are exposed at:

https://<your-edge-hostname>/actuator/prometheus

No authentication is required to scrape this endpoint.

Metrics reference

tb-edge metrics

Metric	Stat names	Description
`attributes_queue_${index}`	`totalMsgs`, `failedMsgs`, `successfulMsgs`	Attribute writes to the database. Multiple queues are used for performance.
`ruleEngine_${queue_name}`	`totalMsgs`, `failedMsgs`, `successfulMsgs`, `tmpFailed`, `failedIterations`, `successfulIterations`, `timeoutMsgs`, `tmpTimeout`	Rule engine message processing, per queue (Main, HighPriority, SequentialByOriginator, etc.).
`ruleEngine_${queue_name}_seconds`	per `tenantId`	Time taken to process messages in each queue.
`core`	`totalMsgs`, `toDevRpc`, `coreNfs`, `sessionEvents`, `subInfo`, `subToAttr`, `subToRpc`, `deviceState`, `getAttr`, `claimDevice`, `subMsgs`	Internal system message processing.
`jsInvoke`	`requests`, `responses`, `failures`	Requests to the JavaScript executors.
`attributes_cache`	`hit`, `miss`	Attribute cache requests.

Rule engine stat descriptions:

Stat	Description
`tmpFailed`	Messages that failed and were reprocessed later
`tmpTimeout`	Messages that timed out and were reprocessed later
`timeoutMsgs`	Messages that timed out and were discarded
`failedIterations`	Processing iterations where at least one message was not processed successfully

Core stat descriptions:

Stat	Description
`toDevRpc`	Processed RPC responses from Transport services
`sessionEvents`	Session events from Transport services
`subInfo`	Subscription info messages from Transport services
`subToAttr`	Subscribe-to-attribute-updates requests from Transport services
`subToRpc`	Subscribe-to-RPC requests from Transport services
`getAttr`	Get-attributes requests from Transport services
`claimDevice`	Device claim requests
`deviceState`	Processed device state changes
`subMsgs`	Processed subscription messages
`coreNfs`	Processed internal system messages

Transport metrics

Metric	Stat names	Description
`transport`	`totalMsgs`, `failedMsgs`, `successfulMsgs`	Requests received by Transport from Core.
`ruleEngine_producer`	`totalMsgs`, `failedMsgs`, `successfulMsgs`	Messages pushed from Transport to the Rule Engine.
`core_producer`	`totalMsgs`, `failedMsgs`, `successfulMsgs`	Messages pushed from Transport to the ThingsBoard node device actor.
`transport_producer`	`totalMsgs`, `failedMsgs`, `successfulMsgs`	Requests from Transport to Core.

PostgreSQL metrics

Metric	Stat names	Description
`ts_latest_queue_${index}`	`totalMsgs`, `failedMsgs`, `successfulMsgs`	Latest telemetry writes to the database. Multiple queues are used for performance.
`ts_queue_${index}`	`totalMsgs`, `failedMsgs`, `successfulMsgs`	Telemetry writes to the database. Multiple queues are used for performance.

Was this helpful?