PMM-15336: fix vanishining qan search and search with whitespace

[fabio-silva]

https://jira.percona.com/browse/PMM-15336

[JNKPercona]

Server docker: perconalab/pmm-server-fb:PR-4545-f56da1b
Client docker: perconalab/pmm-client-fb:PR-4545-f56da1b
Client tarball: https://s3.us-east-2.amazonaws.com/pmm-build-cache/PR-BUILDS/pmm-client/pmm-client-PR-4545-f56da1b.tar.gz
Staging instance: https://pmm.cd.percona.com/job/pmm3-aws-staging-start/parambuild/?DOCKER_VERSION=perconalab/pmm-server-fb:PR-4545-f56da1b&CLIENT_VERSION=https://s3.us-east-2.amazonaws.com/pmm-build-cache/PR-BUILDS/pmm-client/pmm-client-PR-4545-f56da1b.tar.gz

[JNKPercona]

API tests have succeded: https://pmm.cd.percona.com/job/pmm3-api-tests/7164/

[travagliad]

Looked at the red @fb-alerting|@fb-settings check in run 32919785512 — it's not your change. The job never got to running a test: mysql/mysql-setup.yml died in setup with ERROR 2002 (HY000): Can't connect to local MySQL server through socket '/tmp/mysql.sock', because the playbook hard-restarts the MySQL container (SIGKILL after the 10s grace, so mysqld comes back through InnoDB crash recovery) and then waits a fixed 5 seconds before connecting as root — a budget that runs out on a slow runner.

That's a pmm-qa setup bug, and it's the only FB job using --database mysql, which is why nothing else in the run went red. Fix is up as percona/pmm-qa#1254: poll with mysqladmin ping instead of sleeping. Once that merges, re-running the failed job here should get you green.

---

Generated by Claude Code