The contents of this KIP were authored by Jun Rao.

Status

Current state: Under Discussion

Discussion thread:

JIRA: Unable to render Jira issues macro, execution error.

Please keep the discussion on the mailing list rather than commenting on the wiki (wiki discussions get unwieldy fast).

Motivation

Ensuring that the Kafka Controller is healthy is an important part of monitoring the health of a Kafka Cluster. However, the metrics currently exposed are not sufficient for reliably detecting issues like slow progress or deadlocks. We propose a few new metrics that will solve this issue. Even though KAFKA-5028 will potentially fix existing deadlocks, there will still be known (and potentially unknown) issues that can cause slow or no progress so these metrics will still be useful.

Public Interfaces

New metrics in controller:

(1) kafka.controller:type=KafkaController,name=ControllerState:type: Gauge

value: reporting the state the controller is in 
Valid states:
0 - idle
1 - starting
2 - resigning
3 - broker change
4 - topic creation
5 - topic deletion
6 - partition reassigning
7 - auto leader balancing
8 - manual leader balancing
9 - controlled shutdown

(2). kafka.controller:type=ControllerStats,name=ControllerStartRateAndTimeMs
type: timer
value: rate and latency for the controller to start

(3). kafka.controller:type=ControllerStats,name=ControllerResignRateAndTimeMs
type: timer
value: rate and latency for the controller to resign

(4). kafka.controller:type=ControllerStats,name=TopicCreationRateAndTimeMs
type: timer
value: rate and latency for the controller to create new topics

(5). kafka.controller:type=ControllerStats,name=TopicDeletionRateAndTimeMs
type: timer
value: rate and latency for the controller to delete topics

(6). kafka.controller:type=ControllerStats,name=PartitionReassigningRateAndTimeMs
type: timer
value: rate and latency for the controller to reassign partitions

(7). kafka.controller:type=ControllerStats,name=AutoLeaderBalancingRateAndTimeMs
type: timer
value: rate and latency for the controller to auto balance the leaders

(8). kafka.controller:type=ControllerStats,name=ManualLeaderBalancingRateAndTimeMs
type: timer
value: rate and latency for the controller to manually balance the leaders

(9). kafka.controller:type=ControllerStats,name=ControlledShutdownRateAndTimeMs
type: timer
value: rate and latency for the controller to shut down a broker in a controlled way

No need for ControllerStats.BrokerChangeRateAndTimeMs since there is already an existing LeaderElectionRateAndTimeMs.

We also want to know the size of the queue in ControllerChannelManager

(10) kafka.controller:type=ControllerChannelManager,name=TotalQueueSize

(11) kafka.controller:type=ControllerChannelManager,name=QueueSize,brokerId=10

Since quite a few jiras reported continuous errors due to ""Cached zkVersion 54 not equal to that in zookeeper, skip updating ISR". It would be useful to measure the occurrences of failed ISR update in ZK.

(12) kafka.cluster:type=Partition,name=FailedIsrUpdateRate

Proposed Changes

We will add the relevant metric type to one of KafkaController, ControllerStats, ControllerChannelManager or Partition as specified in the Public Interfaces section.

Compatibility, Deprecation, and Migration Plan

We are introducing new metrics so there is no compatibility impact.

Rejected Alternatives

• Don't add these metrics: it's currently difficult to detect these issues, they impact cluster health and the overhead of the proposed metrics is low.