Reduce cache size of CI caches

[hosekpeter]

This pull request introduces a new workflow for cleaning up GitHub Actions caches and includes documentation to explain its purpose and usage. The workflow aims to manage cache growth by removing old and duplicate entries, improving efficiency and reducing disk usage.

Workflow Implementation:

• Added a new workflow, .github/workflows/gh-cache-cleanup.yml, to automate GitHub Actions cache cleanup. It includes features such as scheduled weekly and monthly cleanups, cache deduplication, and a dry-run mode for safe testing.

Documentation:

• Added docs/github-actions-cache-cleanup.md to describe the problem of cache growth, the solution provided by the new workflow, and detailed instructions for configuration, usage, and testing.Jira: XXXXX

Changes:

• ...

---

[hosekpeter]

@jachym-tousek-keboola @Matovidlo take a look: https://github.com/keboola/keboola-as-code/actions/runs/14515804360/job/40724502227?pr=2336

[Matovidlo]

For some reason we stopped using the S3 bucket, now we have GH cache which we do not need to clean at all it has it's own lifecycle.
This cleanup is needed only for S3 bucket this gh cache cleanup is done automatically

[hosekpeter]

@Matovidlo I read somewhere the GitHub cache cleaner can't to remove the duplicate keys. You can set up it in the own workflow. It is not by default.

[Matovidlo]

LGTM. We need to take a look on caching using S3 as this cache is overloaded.
Typically it is 24-48 hours present