Sprint Planning 12-2 to 12-16

[dmarticus]

Team Flags

Support

• 12-2 to 12-9 @dmarticus
• 12-9 to 12-16 @havenbarnes

Retro

Things we got done

• Track feature flag called events in usage reports so we can track up who set up flags successfully @dmarticus
• generally ironing out service (centralizing error handling, reducing memory usage overhead with smarter copying/cloning)
  • found a bug in the following two paths (1) feature flag payloads, and (2) group targeting without group overrides. PR to come later today.
• land v0 of flag staleness feature @havenbarnes
• Fixed issue with feature flag payloads SDK methods on the server-side (Node, Python) not correctly sending feature_flag_response payloads @dmarticus
• Show feature flag payload as official PostHog property @havenbarnes
• Adding some debugging event properties to $feature_flag_called to help indicate when SDKs are using bootstrapping

Things we didn't get done and why

• testing/productionizing new /flags service @dmarticus
  • functional testing via django-compat integration tests
  • "real" testing via mirroring traffic to new service (meeting with ingestion team 11-20 to go through this.

As I was writing more integration tests, I came across some bugs, and needed to fix them.

• Start work on feature management UI @havenbarnes
  • needed realignment on the correct UI abstractions for this feature; made an initial PR to start that conversation, and are iterating on it.

---

Sprint

High priority

• testing/productionizing new /flags service @dmarticus
  • fixing P0 bugs along /flags hotpath
  • functional testing via django-compat integration tests
  • "real" testing via mirroring traffic to new service

Side Quests

• Continue to make sentry more actionable and useful @havenbarnes
• Figure out how to cut off the flags bit of /decide (and, in future /flags) when folks go over – right now we don't stop serving flags when overage happens @dmarticus
• Continue to look at growth metrics and see what other properties we can add to better understand our growth @dmarticus

Things to chew on

• Can we know when people are starting to make new features?

OKRs

1. Ship new feature flags in production and accomplish the 10x scale + 2x speed boost performance targets (ideally decreases infra spend by N% as well, will try to figure out a more specific target going forward)
2. Maintain current /decide service and make optimizations.