* fix(query-performance): avoid mounting systemLogic for the sake of feature flag internals
instance status does some heavy queries so mounting it each time is expensive. Instead lets bundle away the relevant
logic into where it doesnt get loaded needlessly
* fix(query-performance): avoid loading /api/instance_status all the time
This endpoint does some expensive queries not needed for majority of navigation
* Add tests for new endpoint
* Cache whether DLQ is OK for an hour
* Remove redis caching
* Skip DLQ checks on cloud
* Improve a comment
* Simplify feature flag fix
* api
* add performed_rollback
* add celery task and tests
* rollback test
* remove first and last
* add sentry stuff
* basic auto rollback UI
* fix errors
* testable
* add errors rollback ui
* clean up sentry keys
* clean up some ui stuff
* add some sentry context
* update ui
* fix celery
* Update posthog/api/feature_flag.py
Co-authored-by: Neil Kakkar <neilkakkar@gmail.com>
* add sentry instructions when not enabled
* add sentry context
* merge migration
* remove unnecessary field right now and update UI to 7 day trailing average
* fix migration
* fix frontend type
* activity
* reset migratioN'
* remove default
* update test
* add feature flag
* add view for conditions and make sure insight loads
* Update snapshots
Co-authored-by: Neil Kakkar <neilkakkar@gmail.com>
Co-authored-by: Li Yi Yu <li@posthog.com>
Co-authored-by: github-actions <41898282+github-actions[bot]@users.noreply.github.com>
