feat(telemetry): support report event to telemetry #17486
Merged
Task list completed / task-list-completed
Started
2024-07-29 03:25:29
ago
1 / 8 tasks completed
7 tasks still to be completed
Details
Required Tasks
| Task | Status |
|---|---|
| recovery: normal recovery, failure recovery (meet failure in nodes), ad-hoc recovery (request from the frontend) | Incomplete |
| source/sink: report when one executor builds | Incomplete |
| I have written necessary rustdoc comments | Incomplete |
| I have added necessary unit tests and integration tests | Incomplete |
| I have added test labels as necessary. See details. | Incomplete |
| I have added fuzzing tests or opened an issue to track them. (Optional, recommended for new SQL features #7934). | Incomplete |
| My PR contains breaking changes. (If it deprecates some features, please create a tracking issue to remove them in the future). | Incomplete |
All checks passed in ./risedev check (or alias, ./risedev c) |
Completed |
| My PR contains critical fixes that are necessary to be merged into the latest release. (Please check out the details) | Incomplete |
| My PR needs documentation updates. (Please use the Release note section below to summarize the impact on users) | Incomplete |
First, why do we need these get_set defined in different modules, but not just in common? Different nodes will just call it at different places. (Like how we share system params for different nodes) |
Incomplete |
Regardless of what their values are, we just created tracking_id and session_id once in start_telemetry_reporting, so why aren't they singletons? Can you give an example how can we break it? |
Incomplete |
| Using it on individual functions is ok. | Incomplete |
Whether top-level or function-level, #[expect(dead_code, reason="xxx")] is better. Because it tells the reason, and tells you to remove it when the lint is no longer fullfilled. |
Incomplete |
| The tracking ID is unique and immutable across the entire cluster. I support implementing this as a global singleton. | Incomplete |
| The session ID is updated whenever a node is created or restarted. In single-node mode, all nodes run within the same process. Using a singleton in this context might cause these session IDs to overlap, failing to reflect the true state of the cluster. | Incomplete |
| removing the session id part in the events framework, but keep it in stats report, may remove it in future pr. | Incomplete |
| impl the tracking id as a singleton in telemetry_common. | Incomplete |
Loading