PMM-14262: InnoDB redo log space calculation

[mattiasimonato]

https://jira.percona.com/browse/PMM-14262

• PMM-14262: InnoDB redo log space calculation percona/pmm#5970

[JNKPercona]

Server docker: perconalab/pmm-server-fb:PR-4572-fc7f67e
Client docker: perconalab/pmm-client-fb:PR-4572-fc7f67e
Client tarball: https://s3.us-east-2.amazonaws.com/pmm-build-cache/PR-BUILDS/pmm-client/pmm-client-PR-4572-fc7f67e.tar.gz
Staging instance: https://pmm.cd.percona.com/job/pmm3-aws-staging-start/parambuild/?DOCKER_VERSION=perconalab/pmm-server-fb:PR-4572-fc7f67e&CLIENT_VERSION=https://s3.us-east-2.amazonaws.com/pmm-build-cache/PR-BUILDS/pmm-client/pmm-client-PR-4572-fc7f67e.tar.gz

[JNKPercona]

API tests have succeded: https://pmm.cd.percona.com/job/pmm3-api-tests/7416/

[travagliad]

FB Tests red here is not caused by this PR. One job failed — E2E / Percona Server for MySQL UI integration tests / @pmm-ps-integration — and its own test step reported green; the real failures only surfaced in the Launchable gate: PMM-T324 and PMM-T430 in e2e_tests/tests/dashboards/mysql/mysqlDashboards.test.ts, each failing 3/3, so deterministic rather than flaky.

Both are stale pmm-qa expectations left behind by percona/pmm 3dd56df (PMM-15308, "Trim whitespace from dashboard titles", percona/pmm#5767), which merged at 14:49 UTC today — before your perconalab/pmm-server-fb:PR-4572-fc7f67e image was built at 16:17 UTC. The tests still expect "Applied Transactions " with a trailing space and a " " spacer panel that is now empty-titled.

I reproduced both on a throwaway VM against your exact FB image, then confirmed they pass with the fix: percona/pmm-qa#1449. Nothing for you to change here — once that merges, a rebuild or a re-run of the failed job should come back green.

---

Generated by Claude Code

[travagliad]

Follow-up: the fix merged — percona/pmm-qa#1449 is now on main as 1e3511f4.

Re-running the failed @pmm-ps-integration job on run 35367870350 is enough to clear this — no rebuild needed. The job checks out pmm-qa by branch name (ref: main in runner-e2e-tests-playwright.yml), so a re-run picks up the corrected test expectations while keeping your existing perconalab/pmm-server-fb:PR-4572-fc7f67e image, which is the one that should be under test. I've left the re-run to you rather than spending the matrix on your PR.

---

Generated by Claude Code