apache / apache/datafusion

`MetricValue::Custom` encoding\decoding support

Open
#16,945 0 comments 0 reactions 0 assignees View on GitHub
enhancement
Dominant language
Rust
Stars
9.3k
Forks
2.4k
Avg merge
3d 7h
Merged PRs (30d)
344

Description

### Is your feature request related to a problem or challenge?

Datafusion 49 brings `MetricValue::Custom { name: Cow<'static, str>, value: Arc,}` which can be used to extend metric handling. With datafusion ballista shipping execution metrics from executors to scheduler question it is not clear how can custom value can be encoded/decoded more details about metric serialisation can be found at [1] ? At the moment with the effort to bring datafusion 49 support to ballista [2] we would throw error in case of custom metrics value.

[1]: https://github.com/milenkovicm/arrow-ballista/blob/e1e9f6ca423fd558664a7f2fb3b1bc3ed07d7db8/ballista/core/src/serde/scheduler/to_proto.rs#L121
[2]: https://github.com/apache/datafusion-ballista/pull/1285

### Describe the solution you'd like

We need to provide a clean path to support encoding/decoding of `MetricValue::Custom`

### Describe alternatives you've considered

Alternative approach would be current approach where custom metrics value would trigger error or to be ignored (somehow)

### Additional context

_No response_

Contributor guide

Open the contributing guide

Research direction

Start with the linked ballista/core/src/serde/scheduler/to_proto.rs location and the DataFusion 49 MetricValue::Custom definition. Compare the existing metric serialization path with the linked Ballista support work, then define encoding and decoding behavior for custom values, including the fallback behavior. Done means custom metrics no longer cause unsupported-value errors and round-trip behavior is covered.

Written by the indexing model from the issue text.

Assessment

Tech stack
rust
Domain
distributed-systems
Issue type
Feature
Difficulty
4/5
Estimated time
3-5 days
Activity status
Stale
Clarity
Mostly clear
Newbie friendliness
38/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.