Skip to content

[EPIC] Route cpu-intensive grouping tasks to their own queue #54097

Description

@loewenheim

New stacktrace grouping rules released on 2023-06-19 are causing increased CPU load for cocoa and native events specifically. Therefore, we want to route save_event tasks for such events to their own dedicated queue.

### Backend changes
- [ ] https://github.com/getsentry/sentry/pull/54041
- [ ] https://github.com/getsentry/sentry/pull/54071
- [ ] https://github.com/getsentry/sentry/pull/54162
- [ ] https://github.com/getsentry/sentry/pull/54137
- [ ] https://github.com/getsentry/sentry/pull/54163
- [ ] https://github.com/getsentry/sentry/pull/54674
### Ops changes
- [ ] https://github.com/getsentry/ops/pull/7616
- [ ] https://github.com/getsentry/ops/pull/7617

Open questions:

  • What should happen to events that could potentially go to the save_event_highcpu queue or the save_event_attachments queue? Which takes priority?
    • For now, we are going to give priority to attachments. We can still revisit this if there are CPU problems on that queue as well.

Metadata

Metadata

Assignees

Labels

Scope: BackendAutomatically applied to PRs that change backend components
No fields configured for issues without a type.

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions