Set an explicit one-hour fact cache timeout

[ra5alghu1]

The checkout uses a persistent fact cache without an explicit timeout, so Ansible defaults to 24 hours. This sets it to one hour, as suggested in #25.

I kept caching enabled and added instructions for a fresh scan with --flush-cache. I also updated the existing cache notes so the docs agree. Discovery still refreshes its health-related facts on every scan. The setting applies to the source checkout; Galaxy users keep their own configuration.

All 212 tests pass, along with pre-commit, ansible-lint, and the playbook syntax check. No managed-host scan or Docker scenario was run for this configuration and documentation change.

[coderabbitai]

Important

• 🔍 Trigger review

This repository does not receive automatic reviews because it has fewer than 10 stars.

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Advanced

Run ID: 53d2f3fc-4ff6-4656-85cd-696adea46032

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[pr-reviewer-bots]

Thanks for opening this PR. Assigned to @sameeralam3127. Labels added: documentation, size/medium. Checklist: 2/3 passed (see checks tab for details).

[sameeralam3127]

Thanks, @ra5alghu1 — this hits all three acceptance criteria on #25, and I appreciate that you went back and fixed performance.md and runbook.md instead of leaving them describing the old "no expiry" state. That drift is the part these changes usually miss.

I checked the claims rather than taking them on trust:

• ansible.cfg is in galaxy.yml's build_ignore, so "not shipped in the Galaxy collection" is accurate.
• roles/vitals_scan/tasks/discovery.yml does explicitly re-gather min/hardware/network/virtual, and all three playbooks use gather_facts: true with gathering = smart — so the timeout genuinely bounds the implicit gather, which is the hazard the issue described.
• pytest -q → 212 passed, ansible-lint roles/ playbooks/ molecule/ demo/ passed, syntax check passed.

One thing I'd like before merging. CONTRIBUTING asks for a test with every behaviour change, and four docs now quote the one-hour value, so I'd like it pinned: a small test that asserts ansible.cfg sets fact_caching_timeout = 3600. tests/test_ci_floor.py is the pattern to follow — it exists for the same reason, stopping config that docs depend on from drifting silently. configparser on the repo's ansible.cfg is enough; no new fixture needed.

Two minor notes, neither blocking:

• ### Cached facts is the only heading in troubleshooting.md and jumps from # to ###. ## reads better and keeps the same #cached-facts anchor the other three docs link to.
• The intake bot is unhappy only about the title not being Conventional Commits. Ignore it — the repo's own history isn't, and I'll set the final title on merge.

CI needs my approval to run on a first PR; I'll kick that off now.

[sameeralam3127]

Heads-up, @ra5alghu1: I pushed one commit to your branch rather than asking you to rebase. #95 and #96 landed first and all three PRs added an entry to the same [Unreleased] → ### Changed block, so CHANGELOG.md conflicted. The commit is a plain merge of main into your branch keeping all three entries — no force-push, your commits are untouched.

Verified on the merged result before pushing: pytest -q 213 passed, ansible-lint roles/ playbooks/ molecule/ demo/ passed, syntax check passed. The diff against main is still exactly your six files.

I've approved the re-run and this goes in once it's green. On the test I asked for — I'm merging without it and filing a follow-up instead, since the config change itself is sound and I'd rather not hold the fix. Thanks for a careful change; going back to fix performance.md and runbook.md so they stopped describing the old "no expiry" state was the right instinct.

[ra5alghu1]

Thanks for sorting out the merge conflict and re-running the checks. Good catch on the docs drift too. I’ll leave the extra test for the follow-up then. Сб, 26 сент. 2026 г. в 20:50, SAMEER ALAM ***@***.***>:

…

Merged #97 <#97> into main. — Reply to this email directly, view it on GitHub <#97?email_source=notifications&email_token=AT6SPDJO54MMXPMLKMWXUQL5Q7QU5A5CNFSNUABQM5UWIORPF5TWS5BNNB2WEL2JONZXKZKFOZSW45CON52GSZTJMNQXI2LPNYXTGMJYHE3DSMBRGIZTLJTSMVQXG33OU5WWK3TUNFXW5JLFOZSW45FMMZXW65DFOJPWG3DJMNVQ#event-31896901235>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AT6SPDLO7S7WXL7MHENSRAL5Q7QU5AVCNFSNUABGKJSXA33TNF2G64TZHMYTCNJSGM2DKMZWGQ5US43TOVSTWNJVHEZDIOBTHAYTBILWAI> . You are receiving this because you were mentioned.Message ID: ***@***.***>