-
Story
-
Resolution: Unresolved
-
Undefined
-
None
-
None
-
None
-
Product / Portfolio Work
-
False
-
-
False
-
-
-
-
Rox Sprint 4.8G - Global, Rox Sprint 4.8H - Global, Rox Sprint 4.9A - Global, Rox Sprint 4.9B - Global, Rox Sprint 4.9C - Global, Rox Sprint 4.9D - Global, Rox Sprint 4.9E - Global, Rox Sprint 4.9F - Global, Rox Sprint 4.9G - Global, Rox Sprint 4.9H - Global, Rox Sprint 4.10A - Global, Rox Sprint 4.9I - Global
-
0
Automation engineers face various toil related to temporary clusters that are spun up for CI or manual testing for developers.
For example, engineers need to review the provisioned infrastructure in Azure, AWS and GCP accounts on a weekly basis to detect leaked resources.
We have implemented Janitors in the past to clean up orphaned resources, but they struggle with *KS/OCP clusters managed by our infra.rox.systems tool
- https://github.com/stackrox/infra
- https://github.com/stackrox/janitor-aws/
- https://github.com/stackrox/janitor-azure/
- https://github.com/stackrox/janitor
Additionally, after detection, there are no standardized ways to clean up such clusters, so leftovers may remain ever after manual cleanups.
Currently remaining work:
- Refactor GCP and Azure Janitors to Cloud Custodian – the same tool that AWS Janitor uses
- Implementation of H2 and H3 of the Infra Janitor proposal: https://docs.google.com/document/d/1-Krdz86l7uvOwM-BqlAcM09ni3jhiqfjbbDCXUvfolk/edit?tab=t.0
