Increase HelmRelease readiness timeouts that were too short for a
freshly-started QEMU/Talos sandbox:
- qdrant: 60s -> 180s
- redis: 20s -> 120s
- vminstance vm-disk: 5s -> 120s, vm-instance: 5s -> 180s
- mariadb: 30s -> 120s
- clickhouse: 20s -> 120s
- kafka: 30s -> 120s
- mongodb: 60s -> 120s
Add diagnostic output (HelmRelease yaml, pods, events) on failure in
all affected bats files, following the pattern already used in
harbor.bats.
Add cleanup before retry in the E2E workflow: introduce
hack/e2e-cleanup-app.sh and a cleanup-test-app-% Makefile target that
removes leftover CRs and HelmReleases from tenant-test namespace before
each subsequent attempt, so retries do not run against a dirty cluster.
Co-Authored-By: Claude <noreply@anthropic.com>
Signed-off-by: Andrei Kvapil <kvapss@gmail.com>
8ff2ba3337