Summarize Component Ownership Model feedback after pilot rollout

Component Ownership Model Feedback Summary

This issue captures feedback from the pilot rollout of the Component Ownership Model during the Data Insights Platform project. The feedback identifies key areas where the model and supporting infrastructure need improvement to enable teams to own infrastructure components independently.

Key Challenges Identified

1. Observability Integration

  • Metrics labels require SRE coordination
  • High learning curve for jsonnet dashboards
  • Slow review turnaround times
  • Team: Observability
  • Solution in progress: Epic &1711 addressing metrics scraping/export

2. Vault Setup & Access

  • Bootstrapping for new clusters is painful
  • Teams outside infra lack default access
  • CLI access requires ops GKE cluster access
  • Team: Runway
  • Action: Create clear path for non-SREs to set up clusters gitlab-com/gl-infra/platform/runway/team#780

3. Helmfile Onboarding

4. ops.gitlab.net Access

  • Teams lack access to registry and pipelines
  • Limits independent debugging and contribution
  • Team: Runway
  • Action: Epic &1647 proposes moving config-mgmt to gitlab.com with mirroring along with other details

5. Metrics Catalogue Documentation

6. Naming Conventions

  • Inconsistent cluster/resource naming
  • Team: Runway
  • Action: Establish formalized guidelines #28195

7. Component Ownership Model Handbook

  • Needs expansion and accuracy verification
  • Team: Maybe Stephanie or Andrew?
  • Action: Update Component Ownership Model handbook with detailed procedures and validation #28194

Core Issue

The model is sound, but supporting infrastructure and documentation create barriers to independent ownership. Teams still require significant SRE involvement for setup and debugging, contradicting the goal of autonomous component ownership.

Related to: gl-infra#1637 (closed)

For detailed feedback and discussion, see the comments on this issue.

Edited by Stephanie Jackson