Add Kafka Connect as a built‑in JMX metrics target #15561
+1,080
−0
There are no files selected for viewing
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,139 @@ | ||
| # Kafka Connect Metrics | ||
|
|
||
| Here is the list of metrics based on MBeans exposed by Apache Kafka Connect. String-valued JMX | ||
| attributes are exported as `UpDownCounter` metrics with value `1` and only include connector/task | ||
| identifiers alongside any state-mapping attributes. | ||
|
|
||
| ## Compatibility | ||
|
|
||
| This rule set targets both Apache Kafka Connect and Confluent Platform. Apache documents several | ||
| metrics not surfaced in Confluent docs (worker rebalance protocol, per-connector task counts on | ||
| workers, source transaction size stats, and sink record lag max); all of them are included below. Status metrics use | ||
| the superset of values across both variants (connector: running, paused, stopped, failed, | ||
| restarting, unassigned, degraded; task: running, paused, failed, restarting, unassigned, | ||
| destroyed) and fall back to `unknown` for any new values. Differences in bean placeholder | ||
| formatting between the docs are cosmetic; bean names align across both variants. | ||
|
|
||
| ## Worker metrics | ||
|
|
||
| | Metric Name | Type | Unit | Attributes | Description | | ||
| |-----------------------------------------------------------|---------------|-------------|------------|---------------------------------------------------------------------------| | ||
| | kafka.connect.worker.connector.count | UpDownCounter | {connector} | | The number of connectors run in this worker. | | ||
| | kafka.connect.worker.connector.startup.attempts | Counter | {attempt} | | The total number of connector startups that this worker has attempted. | | ||
| | kafka.connect.worker.connector.startup.failure.percentage | Gauge | 1 | | The average percentage of this worker's connectors starts that failed. | | ||
| | kafka.connect.worker.connector.startup.failure.total | Counter | {startup} | | The total number of connector starts that failed. | | ||
| | kafka.connect.worker.connector.startup.success.percentage | Gauge | 1 | | The average percentage of this worker's connectors starts that succeeded. | | ||
| | kafka.connect.worker.connector.startup.success.total | Counter | {startup} | | The total number of connector starts that succeeded. | | ||
|
Comment on lines
+22
to
+26
Contributor
There was a problem hiding this comment. we can simplify this:
I would suggest to do the following:
Author
There was a problem hiding this comment. Noted, will refactor, thanks! |
||
| | kafka.connect.worker.task.count | UpDownCounter | {task} | | The number of tasks run in this worker. | | ||
| | kafka.connect.worker.task.startup.attempts | Counter | {attempt} | | The total number of task startups that this worker has attempted. | | ||
| | kafka.connect.worker.task.startup.failure.percentage | Gauge | 1 | | The average percentage of this worker's tasks starts that failed. | | ||
| | kafka.connect.worker.task.startup.failure.total | Counter | {startup} | | The total number of task starts that failed. | | ||
| | kafka.connect.worker.task.startup.success.percentage | Gauge | 1 | | The average percentage of this worker's tasks starts that succeeded. | | ||
| | kafka.connect.worker.task.startup.success.total | Counter | {startup} | | The total number of task starts that succeeded. | | ||
|
Comment on lines
+28
to
+32
Contributor
There was a problem hiding this comment. we can do the same simplification as with the connector startup metrics. |
||
|
|
||
| ## Worker connector task metrics | ||
|
|
||
| | Metric Name | Type | Unit | Attributes | Description | | ||
| |------------------------------------------------|---------------|--------|-------------------------|----------------------------------------------------------------| | ||
| | kafka.connect.worker.connector.task.destroyed | UpDownCounter | {task} | kafka.connect.connector | The number of destroyed tasks of the connector on the worker. | | ||
| | kafka.connect.worker.connector.task.failed | UpDownCounter | {task} | kafka.connect.connector | The number of failed tasks of the connector on the worker. | | ||
| | kafka.connect.worker.connector.task.paused | UpDownCounter | {task} | kafka.connect.connector | The number of paused tasks of the connector on the worker. | | ||
| | kafka.connect.worker.connector.task.restarting | UpDownCounter | {task} | kafka.connect.connector | The number of restarting tasks of the connector on the worker. | | ||
| | kafka.connect.worker.connector.task.running | UpDownCounter | {task} | kafka.connect.connector | The number of running tasks of the connector on the worker. | | ||
| | kafka.connect.worker.connector.task.total | UpDownCounter | {task} | kafka.connect.connector | The number of tasks of the connector on the worker. | | ||
| | kafka.connect.worker.connector.task.unassigned | UpDownCounter | {task} | kafka.connect.connector | The number of unassigned tasks of the connector on the worker. | | ||
|
Comment on lines
+38
to
+44
Contributor
There was a problem hiding this comment. If those task states are effectively a partition it would be simpler to use a single metric
Author
There was a problem hiding this comment. Ah sorry disregard my previous comment (now deleted), I read your comment too fast. I still think that these are best left as independent metrics as they do provide additional functionality in their form. With your suggested structure, you would lose which tasks is failing. It would be a small group, but some people bind certain tasks to partitions, along with other advanced usage, and having it in this form would provide very useful information in understanding why a task might have failed, or the effects of it failing. |
||
|
|
||
| ## Worker rebalance metrics | ||
|
|
||
| | Metric Name | Type | Unit | Attributes | Description | | ||
| |------------------------------------------------|---------------|-------------|--------------------------------------|---------------------------------------------------------------------------------| | ||
| | kafka.connect.worker.rebalance.completed.total | Counter | {rebalance} | | The total number of rebalances completed by this worker. | | ||
| | kafka.connect.worker.rebalance.protocol | UpDownCounter | 1 | kafka.connect.protocol.state | The Connect protocol used by this cluster. | | ||
| | kafka.connect.worker.rebalance.epoch | UpDownCounter | {epoch} | | The epoch or generation number of this worker. | | ||
| | kafka.connect.worker.rebalance.avg.time | Gauge | s | | The average time in milliseconds spent by this worker to rebalance. | | ||
| | kafka.connect.worker.rebalance.max.time | Gauge | s | | The maximum time in milliseconds spent by this worker to rebalance. | | ||
| | kafka.connect.worker.rebalance.active | UpDownCounter | 1 | kafka.connect.worker.rebalance.state | Whether this worker is currently rebalancing. | | ||
| | kafka.connect.worker.rebalance.since_last | Gauge | s | | The time in milliseconds since this worker completed the most recent rebalance. | | ||
|
|
||
| ## Connector metrics | ||
|
|
||
| Attributes: `kafka.connect.connector` and the state attribute shown. | ||
|
|
||
| | Metric Name | Type | Unit | Attributes | Description | | ||
| |---------------------------------|---------------|------|----------------------------------------|-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------| | ||
| | kafka.connect.connector.type | UpDownCounter | 1 | kafka.connect.connector.type | The type of the connector. One of 'source' or 'sink'. | | ||
| | kafka.connect.connector.status | UpDownCounter | 1 | kafka.connect.connector.state | Connector lifecycle state indicator (1 when the state matches the attribute value); accepts running, paused, stopped, failed, restarting, unassigned, degraded, or unknown. | | ||
|
|
||
| ## Connector task metrics | ||
|
|
||
| All metrics include `kafka.connect.connector` and `kafka.connect.task.id`. Attributes column lists any additional state attributes. | ||
|
|
||
| | Metric Name | Type | Unit | Attributes | Description | | ||
| |-----------------------------------------------------|---------------|----------|---------------------------------------------------|--------------------------------------------------------------------------------------------------------------------| | ||
| | kafka.connect.task.batch.size.avg | Gauge | {record} | | The average number of records in the batches the task has processed so far. | | ||
| | kafka.connect.task.batch.size.max | Gauge | {record} | | The number of records in the largest batch the task has processed so far. | | ||
| | kafka.connect.task.connector.type | UpDownCounter | 1 | kafka.connect.task.connector.type | The type of the connector. One of 'source' or 'sink'. | | ||
| | kafka.connect.task.offset.commit.avg.time | Gauge | s | | The average time in milliseconds taken by this task to commit offsets. | | ||
| | kafka.connect.task.offset.commit.failure.percentage | Gauge | 1 | | The average percentage of this task's offset commit attempts that failed. | | ||
| | kafka.connect.task.offset.commit.max.time | Gauge | s | | The maximum time in milliseconds taken by this task to commit offsets. | | ||
| | kafka.connect.task.offset.commit.success.percentage | Gauge | 1 | | The average percentage of this task's offset commit attempts that succeeded. | | ||
| | kafka.connect.task.pause.ratio | Gauge | 1 | | The fraction of time this task has spent in the pause state. | | ||
| | kafka.connect.task.running.ratio | Gauge | 1 | | The fraction of time this task has spent in the running state. | | ||
| | kafka.connect.task.status | UpDownCounter | 1 | kafka.connect.task.state | The status of the connector task; supports running, paused, failed, restarting, unassigned, destroyed, or unknown. | | ||
|
|
||
| ## Sink task metrics | ||
|
|
||
| Attributes: `kafka.connect.connector`, `kafka.connect.task.id`. | ||
|
|
||
| | Metric Name | Type | Unit | Attributes | Description | | ||
| |---------------------------------------------------|---------------|-------------|------------|--------------------------------------------------------------------------------------------------------------------------------------| | ||
| | kafka.connect.sink.offset.commit.completion.rate | Gauge | {commit}/s | | The average per-second number of offset commit completions that were completed successfully. | | ||
| | kafka.connect.sink.offset.commit.completion.total | Counter | {commit} | | The total number of offset commit completions that were completed successfully. | | ||
| | kafka.connect.sink.offset.commit.seq | UpDownCounter | {sequence} | | The current sequence number for offset commits. | | ||
| | kafka.connect.sink.offset.commit.skip.rate | Gauge | {commit}/s | | The average per-second number of offset commit completions that were received too late and skipped/ignored. | | ||
| | kafka.connect.sink.offset.commit.skip.total | Counter | {commit} | | The total number of offset commit completions that were received too late and skipped/ignored. | | ||
| | kafka.connect.sink.partition.count | UpDownCounter | {partition} | | The number of topic partitions assigned to this task. | | ||
| | kafka.connect.sink.put.batch.avg.time | Gauge | s | | The average time taken by this task to put a batch of sinks records. | | ||
| | kafka.connect.sink.put.batch.max.time | Gauge | s | | The maximum time taken by this task to put a batch of sinks records. | | ||
| | kafka.connect.sink.record.active.count | UpDownCounter | {record} | | The number of records that have been read from Kafka but not yet completely committed/flushed/acknowledged by the sink task. | | ||
| | kafka.connect.sink.record.active.count.avg | Gauge | {record} | | The average number of records that have been read from Kafka but not yet completely committed/flushed/acknowledged by the sink task. | | ||
| | kafka.connect.sink.record.active.count.max | Gauge | {record} | | The maximum number of records that have been read from Kafka but not yet completely committed/flushed/acknowledged by the sink task. | | ||
| | kafka.connect.sink.record.lag.max | Gauge | {record} | | The maximum lag in terms of number of records that the sink task is behind the consumer's position for any topic partitions. | | ||
| | kafka.connect.sink.record.read.rate | Gauge | {record}/s | | The average per-second number of records read from Kafka for this task before transformations are applied. | | ||
| | kafka.connect.sink.record.read.total | Counter | {record} | | The total number of records read from Kafka by this task since it was last restarted. | | ||
| | kafka.connect.sink.record.send.rate | Gauge | {record}/s | | The average per-second number of records output from the transformations and sent/put to this task. | | ||
| | kafka.connect.sink.record.send.total | Counter | {record} | | The total number of records output from the transformations and sent/put to this task since it was last restarted. | | ||
|
|
||
| ## Source task metrics | ||
|
|
||
| Attributes: `kafka.connect.connector`, `kafka.connect.task.id`. | ||
|
|
||
| | Metric Name | Type | Unit | Attributes | Description | | ||
| |----------------------------------------------|---------------|------------|------------|-------------------------------------------------------------------------------------------------------------| | ||
| | kafka.connect.source.poll.batch.avg.time | Gauge | s | | The average time in milliseconds taken by this task to poll for a batch of source records. | | ||
| | kafka.connect.source.poll.batch.max.time | Gauge | s | | The maximum time in milliseconds taken by this task to poll for a batch of source records. | | ||
| | kafka.connect.source.record.active.count | UpDownCounter | {record} | | The number of records that have been produced by this task but not yet completely written to Kafka. | | ||
| | kafka.connect.source.record.active.count.avg | Gauge | {record} | | The average number of records that have been produced by this task but not yet completely written to Kafka. | | ||
| | kafka.connect.source.record.active.count.max | Gauge | {record} | | The maximum number of records that have been produced by this task but not yet completely written to Kafka. | | ||
| | kafka.connect.source.record.poll.rate | Gauge | {record}/s | | The average per-second number of records produced/polled (before transformation) by this task. | | ||
| | kafka.connect.source.record.poll.total | Counter | {record} | | The total number of records produced/polled (before transformation) by this task. | | ||
| | kafka.connect.source.record.write.rate | Gauge | {record}/s | | The average per-second number of records written to Kafka for this task. | | ||
| | kafka.connect.source.record.write.total | Counter | {record} | | The number of records output written to Kafka for this task. | | ||
| | kafka.connect.source.transaction.size.avg | Gauge | {record} | | The average number of records in the transactions the task has committed so far. | | ||
| | kafka.connect.source.transaction.size.max | Gauge | {record} | | The number of records in the largest transaction the task has committed so far. | | ||
| | kafka.connect.source.transaction.size.min | Gauge | {record} | | The number of records in the smallest transaction the task has committed so far. | | ||
|
|
||
| ## Task error metrics | ||
|
|
||
| Attributes: `kafka.connect.connector`, `kafka.connect.task.id`. | ||
|
|
||
| | Metric Name | Type | Unit | Attributes | Description | | ||
| |-----------------------------------------------------------|---------|-----------|------------|---------------------------------------------------------------| | ||
| | kafka.connect.task.error.deadletterqueue.produce.failures | Counter | {failure} | | The number of failed writes to the dead letter queue. | | ||
| | kafka.connect.task.error.deadletterqueue.produce.requests | Counter | {request} | | The number of attempted writes to the dead letter queue. | | ||
| | kafka.connect.task.error.last.error.timestamp | Gauge | s | | The epoch timestamp when this task last encountered an error. | | ||
| | kafka.connect.task.error.total.errors.logged | Counter | {error} | | The number of errors that were logged. | | ||
| | kafka.connect.task.error.total.record.errors | Counter | {record} | | The number of record processing errors in this task. | | ||
| | kafka.connect.task.error.total.record.failures | Counter | {record} | | The number of record processing failures in this task. | | ||
| | kafka.connect.task.error.total.records.skipped | Counter | {record} | | The number of records skipped due to errors. | | ||
| | kafka.connect.task.error.total.retries | Counter | {retry} | | The number of operations retried. | | ||
Oops, something went wrong.
Oops, something went wrong.
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Is this the number of currently active connectors on this worker ?
Can we derive this value from the startup attempts ?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Yes this is the number of currently active workers. I would avoid having it be derived from startup attempts as depending on the kafka connect configuration, there can be retries. I think this should be its own metric as it is very important. In fact, I would say we can get rid of the startup and have this metric if you want to get rid of one of the two.