[FLINK-13550] Support for CPU FlameGraphs in web UI

• Type: New Feature
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 1.13.0
• Component/s: Runtime / REST, Runtime / Web Frontend
• Labels: pull-request-available
• Release Note: Flink now offers Flamegraphs for each node in the job graph. Please enable this experimental feature by setting the respective configuration flag rest.flamegraph.enabled.
• Assignee: Alexander Fedulov
• Reporter: David Morávek
• Created: 02/Aug/19 08:17
• Updated: 30/Nov/22 11:37
• Resolved: 01/Apr/21 11:47

Description

For a better insight into a running job, it would be useful to have ability to render a CPU flame graph for a particular job vertex.

Flink already has a stack-trace sampling mechanism in-place, so it should be straightforward to implement.

This should be done by implementing a new endpoint in REST API, which would sample the stack-trace the same way as current BackPressureTracker does, only with a different sampling rate and length of sampling.

Here is a little demo of the feature.

Issue Links

• causes: FLINK-30250 The flame graph type is wrong
• is related to: FLINK-22527 Give friendly hint when click FlameGraph without rest.flamegraph.enabled
• relates to: FLINK-22235 Document stability concerns of flamegraphs for heavier jobs
• relates to: FLINK-22237 Flamegraph component calls draw method from template
• relates to: FLINK-22238 Investigate angular native flamegraph library
• relates to: FLINK-22236 Document opt-in behavior of flamegraph
• links to: GitHub Pull Request #15054
• mentioned in: Page Loading...