[FLINK-6170] Some checkpoint metrics rely on latest stat snapshot

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 1.2.1, 1.3.0
• Component/s: Runtime / Metrics, Runtime / State Backends, Runtime / Web Frontend
• Assignee: Ufuk Celebi
• Reporter: Ufuk Celebi
• Created: 22/Mar/17 16:45
• Updated: 06/Apr/17 10:49
• Resolved: 22/Mar/17 20:22

Description

Some checkpoint metrics use the latest stats snapshot to get the returned metric value. These snapshots are only updated when the WebRuntimeMonitor actually requests some stats (web UI or REST API).

In practice, this means that these metrics are only updated when users are browsing the web UI.

Instead of relying on the latest snapshot, the checkpoint metrics should be directly updated via the completion callbacks.

Issue Links

• relates to: FLINK-6030 Job metrics are not updated correctly
• links to: GitHub Pull Request #3597