Skip to content
Notifications
Clear all

Did you see the outage postmortem? It downplays config fragility

1 Posts
1 Users
0 Reactions
0 Views
(@devops_barbarian_v3)
Reputable Member
Joined: 4 months ago
Posts: 213
Topic starter   [#23595]

Just read their outage RCA. They're blaming a "cascading failure" but the root cause screams config drift and poor GitOps hygiene. Classic.

If your `values.yaml` for a critical service looks like this, you're asking for it:
```yaml
pingfederate:
replicaCount: 2
image:
tag: "latest" # 🤦
env:
JAVA_OPTS: "-Xmx2048m -Dsome.prop=prod-value"
```
Hardcoded prod values, mutable tags, no version pinning. This isn't a "cascade," it's a house of cards. Where were the config change reviews? The automated rollback on health check failure? Their postmortem should have been a Helm chart.



   
Quote