Flag default flips on missing config: symptoms, cause, and the test that catches it
A missing or unreadable flag config falls back to the wrong default.
Flag default flips on missing config is documented by 2 accepted public reports from 2 organisations, 2 of them published in the last 5 years.
- Likelihood: low (2 public reports in 5 years)
- Impact: low (2 of 2 reports: degraded)
Symptoms
User-visible
- Function execution was degraded for about 60 minutes.[1]
- Run scheduling was degraded for about 125 minutes.[1]
- A subset of workflow steps failed, and some outbound fetch requests from steps returned empty responses.[1]
- EU customers found the chart formula feature in event segmentation incorrectly gated.[2]
Root causes
- stage: third-party Connectivity to the external feature flag service degraded, so services could not fetch flag values at startup or through streaming updates.[1]
- stage: config Containers started by the new release fell back to default flag values. The defaults moved a subset of internal traffic from the RPC path to a legacy HTTP path and misconfigured another subsystem.[1]
- stage: config An update gated the chart formula feature, and the EU region had no feature flag for that update. EU customers therefore received the gated state.[2]
Public examples
- [1]Inngest, 2026-07-23. During an Inngest release on 10 July 2026, newly started containers could not fetch feature flag values from a third-party flag service and fell back to default configuration, which moved internal traffic onto a legacy HTTP path, overloaded the internal network and degraded function execution and run scheduling until the release was rolled back. primary Root Cause Analysis: July 2026 Service Incidents
- [2]Amplitude, 2023-07-28. An Amplitude update gated the chart formula feature, and because the EU feature flag for it was missing, EU customers fell back to the gated state for about three days. primary EU Event Segmentation Chart Formulas gated unintentionally
Test classes that catch it
| Test class | Stage | Limitation | Basis |
|---|---|---|---|
| Configuration and environment tests | pre-deploy | A configuration test can confirm that every region has each flag and that defaults match production, but it does not measure the load the legacy HTTP path put on the internal network once defaults moved traffic onto it. | by design analysis |
| Chaos and resilience tests | pre-deploy | A chaos test that cuts the flag service exercises the startup fallback, but it does not find a flag that was never created for the EU region, because that region had no dependency failure to inject. | by design analysis |
| Canary and progressive delivery checks | post-deploy | A canary that receives no EU traffic does not see the gated chart formula feature, and a gated feature returns no error for the canary metrics to compare. | by design analysis |
Risk
- Likelihood: low (2 public reports in 5 years)
- Impact: low (2 of 2 reports: degraded)
| Accepted public reports | 2 |
|---|---|
| Published in the last 5 years | 2 |
| Impact stated: degraded | 2 |
Pre-release checklist
- PR-01
- PR-02
- PR-03
- PR-04
- PR-05
- PR-06
- PR-07
- PR-08
First 30 minutes
- FM-01
- FM-02
- FM-03
- FM-04
- FM-05
- FM-06
- FM-07
- FM-08
Download the checklist
What catches this: catalogue entries
Listed because their own documentation or site names configuration and environment tests, chaos and resilience tests, canary and progressive delivery checks on the verification date. Alphabetical. Not a recommendation.
Tools whose documentation names these test classes
- Argo Rollouts canary source verified 2026-10-21
- AWS Fault Injection Service chaos source verified 2026-10-21
- Chaos Mesh chaos source verified 2026-10-21
- Chaos Toolkit chaos source verified 2026-10-21
- Flagger canary source verified 2026-10-21
- Flipt config source verified 2026-10-21
- Grafana k6 chaos source verified 2026-10-21
- Gremlin chaos source verified 2026-10-21
- GrowthBook config source verified 2026-10-21 canary source verified 2026-10-21
- Hoverfly chaos source verified 2026-10-21
- Kayenta canary source verified 2026-10-21
- kube-monkey chaos source verified 2026-10-21
- LaunchDarkly config source verified 2026-10-21 canary source verified 2026-10-21
- LitmusChaos chaos source verified 2026-10-21
- PostHog config source verified 2026-10-21
- Pumba chaos source verified 2026-10-21
- Spinnaker canary source verified 2026-10-21
- Toxiproxy chaos source verified 2026-10-21
- Unleash config source verified 2026-10-21 canary source verified 2026-10-21
- WireMock chaos source verified 2026-10-21
Providers whose site names these test classes
- QualityKiosk Technologies chaos source verified 2026-10-21
- Testhouse chaos source verified 2026-10-21
Only on this page
Flag default flips on missing config in report [1] needed a dependency outage and a release at the same time. Only containers started by the new release fell back to default flag values, because connectivity to the external flag service had degraded [1]. The defaults moved part of internal traffic onto a legacy HTTP path, and DNS query volume rose roughly tenfold [1]. A separate code regression in the same release had started customer impact at 16:10 UTC, and flag initialization errors were observed at 16:19 UTC [1].
Report [2] shows the regional form of the fault, with no outage involved. The EU region had no flag for an update that gated a chart formula feature, so EU customers received the gated state for about three days [2]. A follow-up in report [1] switched a subset of feature flags to static values.
Related incidents
- Rate limit misconfigured A rate limit change blocks legitimate traffic or stops protecting an endpoint.
- Environment mismatch: production points to staging A production service reads a staging endpoint, bucket or database after a config change.
- Flag enabled for wrong segment A feature flag targets the wrong users, region or percentage after a rule edit.
- Config change without restart semantics A config edit is applied to some instances and not others, producing inconsistent behaviour.
- Autoscaling policy misconfigured A scaling change fails to add capacity or scales to zero under load.
Questions
What causes flag default flips on missing config?
In report [1], new containers could not reach the external flag service and fell back to default flag values. In report [2], the EU region had no flag for an update that gated a chart formula feature.
What are the symptoms of flag default flips on missing config?
Function execution was degraded for about 60 minutes and run scheduling for about 125 minutes [1]. EU customers found a chart formula feature incorrectly gated for about three days [2].
Which test class catches flag default flips on missing config, and what does it miss?
Configuration tests, chaos and resilience tests and canary checks are listed, all with the by-design label. A chaos test that cuts the flag service does not find a flag that was never created for the EU region.
Which public reports document flag default flips on missing config?
Two primary reports from two organisations document it. They were published on 2023-07-28 [2] and 2026-07-23 [1].
How long did it take to detect flag default flips on missing config in the public reports?
Neither report states a time to detect. In report [1], customer impact began at 16:10 UTC and flag initialization errors were observed at 16:19 UTC.
Should a team roll back or fix forward when flag default flips on missing config happens?
Report [1] rolled the release back and later switched a subset of flags to static values. Report [2] states that no data was lost and no customer action was required.
Release Incident Playbook (2026). Flag default flips on missing config: symptoms, cause, and the test that catches it, content version 1.0.0. https://releasefailures.com/incidents/flag-default-flips-on-missing-config/. Accessed 2026-10-21.