Expand description
Why a replica did not come up.
A replica that fails its rollout is deleted at once and the controller
tries again, so by the time anyone asks for its logs the instance is
gone (or dying, and incus answers the console request with an error).
The controller therefore reads the failed instance’s output before it
deletes it, puts the last lines in the failure message (the deployment
log, the service’s status) and keeps them here for app_logs, which
shows them beside the live replicas’ logs.
Structs§
- Failed
Attempt - The output of the last replica of a service that failed to come up.
- Failures
- The last failed attempt of each (stack, service).
Functions§
- explain
- The failure of replica
namewith its last output in the message, and the attempt to keep. Called before the instance is deleted. - one_
line - The last
nnon-empty lines oftext, joined with|and cut tomaxcharacters (from the front: the end is what explains it). - replica_
logs - Recent output of a service’s replicas (or one slot’s), by instance. A replica that is being replaced as this is asked, or whose output cannot be read, shows why in place of its text instead of failing the call.