v1.8.0
LatestNew
- challenge: set up several faults together, and keep them unnamed
- challenges: on-call-shift, two faults at once with no names
- go-api: an upstream, a slow start and per-request CPU work, all off
- incidents: cpu-throttling, a CPU limit set to save cost
- incidents: dns-egress-blocked, a security policy that forgot port 53
- incidents: liveness-probe-kill, a probe path the app does not serve
- incidents: missing-secret-key, a rotated key no new pod can find
- incidents: pending-unschedulable, replicas no node can hold
Fixed
- challenge: a silent challenge's failing submit names no fault
- cli: a followed run no longer loses output it wrote just before it ended
- cluster-upgrade-drill: the replacement node gets every build of the lab's apps
- dns-egress-blocked: grade only an app that really calls its upstream
- toolchain: a cancelled run ends when its script does, and leaves nothing behind