Flag default flips on missing config: pre-release and first 30 minutes checklists

A missing or unreadable flag config falls back to the wrong default.

The pre-release checklist maps each item to a test class from the incident page. The first-30-minutes checklist covers detection, containment, the rollback decision and customer communication.

8 pre-release items · 8 first-30-minutes items · verified 2026-10-21

Pre-release checklist

PR-01Compare the flag set of every region with the release's flag list and fail the release when a region such as the EU lacks a flag the update depends on. config
PR-02Read the default value of each flag in the release and confirm it equals the current production value. config
PR-03Confirm the flags that route internal traffic default to the RPC path and not to the legacy HTTP path. config
PR-04Start new containers with the external flag service blocked and confirm they keep the last known production flag values. chaos
PR-05Drop the flag streaming connection on running containers and confirm their flag values stay unchanged. chaos
PR-06Record DNS query volume and connection setup time during the blocked-flag-service test and fail the test when either rises above its baseline. chaos
PR-07Halt the rollout when any canary container logs a feature flag initialization error. canary
PR-08Place a canary in each region, including the EU, and compare the share of EU sessions that use the chart formula feature between canary and baseline. canary

First 30 minutes

FM-01Search container startup logs for feature flag initialization errors and note which release started those containers.
FM-02Test connectivity from the new containers to the external flag service.
FM-03Compare the flag values the new containers loaded with production values, starting with flags that route internal traffic.
FM-04Roll back the release so containers running on default flag values stop serving traffic.
FM-05Confirm DNS query volume, DNS lookup latency and connection setup times return to their baselines after the rollback.
FM-06If one region lacks a flag, create that flag with its production value for the region instead of redeploying.
FM-07Pin the flags that must not change to static values until the flag service is reachable again.
FM-08Post a status update that names the degraded functions or gated feature, the affected region and whether customers need to replay events.

Download the checklist

Pre-release checklist (Markdown)1.7 KB · v1.0.0
First 30 minutes checklist (Markdown)1.2 KB · v1.0.0

Only in this checklist

Every test class in this checklist also appears in at least one sibling checklist in the group; the item texts differ: Compare the flag set of every region with the release's flag list and fail the release when a region such as the EU lacks a flag the update depends on.

Questions

What should be checked before a release to prevent Flag default flips on missing config?

8 pre-release items, PR-01 to PR-08. PR-01: Compare the flag set of every region with the release's flag list and fail the release when a region such as the EU lacks a flag the update depends on.

Is there a rollback plan for Flag default flips on missing config?

The first-30-minutes checklist has 8 items. FM-01: Search container startup logs for feature flag initialization errors and note which release started those containers.

What should a team do in the first 30 minutes after Flag default flips on missing config is detected?

The first-30-minutes checklist has 8 items. FM-01: Search container startup logs for feature flag initialization errors and note which release started those containers.

Which configuration checks apply to Flag default flips on missing config?

PR-01: Compare the flag set of every region with the release's flag list and fail the release when a region such as the EU lacks a flag the update depends on. PR-02: Read the default value of each flag in the release and confirm it equals the current production value. PR-03: Confirm the flags that route internal traffic default to the RPC path and not to the legacy HTTP path.

Is a Flag default flips on missing config checklist still needed if deployment is automated?

Yes. Automation runs the steps; the checklist names what the tests must cover: configuration and environment tests, chaos and resilience tests and canary and progressive delivery checks.

Which pipeline stages does the Flag default flips on missing config checklist cover?

The pre-release items map to the pre-deploy and post-deploy stages.