Clean up disk for "More Go tests"

[dimitarvdimitrov]

TestCompaction_CompactWithSplitting is consistently failing on my branches with preallocate: no space left on device

example https://github.com/grafana/mimir-prometheus/actions/runs/20029126858/job/57433565471#step:6:450

Which issue(s) does the PR fix:

Does this PR introduce a user-facing change?

NONE

---

Note

Upgrade promci for test_go and enable disk cleanup in test_go_more to mitigate space issues.

• CI (.github/workflows/ci.yml):
  • test_go:
    • Bump prometheus/promci to v0.5.3.
    • Remove unused setup_environment NPM options.
  • test_go_more:
    • Add clean-runner-disk: true to setup_environment to free disk space.

^{Written by Cursor Bugbot for commit 197ada7. This will update automatically on new commits. Configure here.}

[Copilot]

Pull request overview

This PR addresses disk space issues in the "More Go tests" CI workflow by enabling disk cleanup before test execution. The change specifically targets failures of TestCompaction_CompactWithSplitting which consistently fails with "preallocate: no space left on device" errors in this fork.

• Enables the clean-runner-disk parameter for the promci setup_environment action
• Adds explanatory comments noting this is a fork-specific fix not present upstream

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

[nitpick] The comment "For some reason a compactor test fails by running out of disk" is vague. Consider being more specific about which test is failing. Based on the PR description, this should reference TestCompaction_CompactWithSplitting.

Suggested improvement:

# TestCompaction_CompactWithSplitting fails by running out of disk space.

Suggested change

	# For some reason a compactor test fails by running out of disk.
	# TestCompaction_CompactWithSplitting fails by running out of disk space.

[dimitarvdimitrov]

welp, this doesn't fix it

[cursor]

Bug: Version mismatch causes clean-runner-disk option to fail

The test_go_more job uses prometheus/[email protected] but attempts to use the clean-runner-disk: true parameter with setup_environment. The test_go job was updated to use promci v0.5.3, which likely introduced the clean-runner-disk feature. Since test_go_more still references v0.4.7, the clean-runner-disk parameter may not be recognized or functional, explaining why the author's comment says "welp, this doesn't fix it". The promci version at line 39 needs to match v0.5.3 for the new option to work.