[FLINK-4202] Add JM metric which shows the restart duration

• Type: Improvement
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 1.1.0
• Fix Version/s: 1.1.0
• Component/s: Runtime / Coordination
• Assignee: Till Rohrmann
• Reporter: Till Rohrmann
• Created: 12/Jul/16 14:14
• Updated: 22/Jul/16 12:19
• Resolved: 22/Jul/16 12:19

Description

It is convenient for users to have a metric which tells you how long the restarting of a job has taken.

I propose a to introduce a Gauge which returns the time between JobStatus.RESTARTING and JobStatus.RUNNING. If the job was not restarted (initial run), then this metric will return -1.

Issue Links

• links to: GitHub Pull Request #2271
• links to: GitHub Pull Request #2284