Add metrics Gauges to track queue backlog in Cluster Executors

Description

Encountered a scenario where the Reconnection executor becomes backlogged because reconnection attempts to hosts were delayed or timing out.

It would be useful for debugging to have Gauges on all Executor's queue sizes in Cluster so we can detect when they become backlogged.

Environment

None

Pull Requests

None

Activity

Show:
Andy Tolbert
March 18, 2015, 7:34 PM

Adds 4 new metrics:

  1. executor-queue-depth

  2. blocking-executor-queue-depth

  3. reconnection-scheduler-task-count

  4. task-scheduler-task-count

These can be used for monitoring whether or not a Cluster's executors are becoming backlogged which could help understand abnormal behavior of the driver.

Fixed

Assignee

Andy Tolbert

Reporter

Andy Tolbert

Labels

PM Priority

None

Affects versions

None

Fix versions

Pull Request

None

Doc Impact

None

Size

None

External issue ID

None

External issue ID

None

Components

Priority

Major
Configure