Platform team weekly
Facilitator: Priya Raman · Date: 2026-09-28 · Note taker: Tomas Vidal
Last week's action items
Wins
- Northwind's SSO migration went live with zero support tickets (Jun and Amara)
- Build times dropped from 14 to 9 minutes after the cache change (Tomas)
- Lena from Support thanked the team for the faster incident status page updates
Metrics
Deploys rose because the faster build makes smaller changes cheap; failure rate fell with them, which is the pattern we want.
Priorities this week
- Finish the Postgres 16 upgrade on the two remaining replicas
- Close the Fernhill audit findings before their Oct 6 review
- Draft the Q4 reliability goals for Thursday's planning
Blockers
| Blocker | Impact | Needs | Owner |
|---|---|---|---|
| No vault admin access for credential rotation | Rotation slips, audit finding stays open | Security to grant temporary role | Jun |
| Replica upgrade window conflicts with a sales demo | Upgrade delayed a week | Move the demo or the window | Priya |
Discussion
- Pager load on weekends (10 min, raised by Amara): four weekend pages in September, three from the same flaky disk alert. Proposal: raise the threshold and add a runbook link.
- Hiring loop (5 min, raised by Priya): two onsite slots needed next week.
Decisions log
| Decision | Why | Decided by | Date |
|---|---|---|---|
| Raise the disk alert threshold to 90% for 30 minutes | Three false pages in a month, no real incident behind them | Team, Priya to confirm | 2026-09-28 |
| Move the replica upgrade to Wednesday night | Keeps the Lumen Labs demo stable on Tuesday | Priya | 2026-09-28 |