Backend Search: Serverless Implementations

[joe-elliott]

• [ ] Code improvements/bug fixes
  • [x] Support pools of external search endpoints
  • [x] Build/Test serverless folder as part of CI
  • [x] tempo-cli command to do backend search
  • [x] Have queriers prefer themselves to the external search endpoints
  • [x] Fix min/max time range on blocks
  • [x] Support searching on fields besides strings
  • [x] Add ingestion time range limits to spans
  • [ ] Improve querier/query-frontend relationship. Can we have per endpoint max_parrallelism?
• [ ] Build tooling to create code packages for all major cloud providers
  • [x] GCP
  • [x] AWS
  • [ ] Azure
• [ ] Create guides for deploying to the major cloud providers
  • [x] GCP
  • [x] AWS
  • [ ] Azure

[yvrhdn]

Determine why coldstart is faster than reusing existing function instances

I wonder if this might be a garbage collection issue? After the execution the function is frozen and garbage collection can not run. So if we are allocating a lot, it will probably have to run during the next execution.

[joe-elliott]

Removed

Determine why coldstart is faster than reusing existing function instances

b/c I no longer think this is an issue with Cloud Functions. I believe I was exhausting a resource in the querier that was causing timeouts and retries.

[github-actions[bot]]

This issue has been automatically marked as stale because it has not had any activity in the past 60 days. The next time this stale check runs, the stale label will be removed if there is new activity. The issue will be closed after 15 days if there is no new activity. Please apply keepalive label to exempt this Issue.

[joe-elliott]

This was completed except for Azure. Splitting off a new issue and closing this.

See #1898