Backstage plugin inventory - which plugins to include in v1

[arjunmehta-git]

Before we configure our Backstage instance, we need to decide which plugins make the cut for v1. Starting list:

Core (always include):

• Software catalog
• TechDocs
• GitHub Actions integration
• Cost Insights
• Search

Strong candidates:

• Kubernetes plugin (show cluster state from catalog)
• PagerDuty plugin (on-call from service page)
• SonarQube plugin (code quality on service page)
• Lighthouse plugin (web performance auditing)

Defer to v2:

• Permission framework (complex setup)
• Scaffolder custom actions (needs testing)
• Bazaar (inner-sourcing - not a priority yet)

Feedback welcome. Let's lock the list by Friday.

[miacycle]

Strong +1 on Kubernetes plugin - devs constantly ask 'where is my service running and is it healthy?' The plugin answers that directly from the catalog page. This is probably our #1 DX win.

One to add: the tech-insights plugin for tracking adoption of golden paths. If we're building golden paths, we should be able to measure how many services actually use them.

[sarajkrishnasingh]

Agree on Kubernetes plugin. On Cost Insights - it requires significant backend work to wire up AWS cost data. I'd defer it to v2 unless we have a clear data source already. The effort/reward ratio is poor for v1.

[arjunmehta-git]

Fair point on Cost Insights - moving it to v2. Tech Insights is a good addition from Mia - adding to strong candidates. @pontusringblom what's your take on PagerDuty vs a simpler on-call display? Do our clients universally use PD?

[pontusringblom]

Not universal - about 60% of clients use PagerDuty. The other 40% use OpsGenie or custom on-call. I'd make PD optional (installed but not enabled by default) and document how to wire up OpsGenie as an alternative. That way we don't block on a vendor decision.