Task · SKT-0245

Stabilise the owned Fleet HA interruption and recovery test under load

Description

TestFleetUnregisterOwnedHAAuthorityInterruptionAndRecovery (cmd/synthkit/fleet_owned_test.go:579) failed once on the remote gate (gaterun 20261026T163726Z-synthkit-15cb7d7f) on a change that does not touch cmd/synthkit: recovery exit status 1 with teardown receipt status remaining (1 of 2 retired). The same code passed the gate an hour earlier. The test uses a deliberate paced 400 ms / 1 s client setting for the client-go rate limiter. Find why recovery can leave a collector unretired on a loaded runner, without weakening assertions or bounds.

Definition of Done

View the source file on GitHub