Missing chapter in your CI playbook: database guardrails
summary
Demonstrates catching database failures inside CI by extracting execution plans and tracing SQL with OpenTelemetry spans, surfacing issues that unit tests miss: a join that pulls 300,000 rows and causes 25s memory churn, a CTE vs non‑CTE query on 8 million rows that runs 13s vs 8s, and schema changes that can rewrite tables for minutes. Integrates into pipelines and GitHub Actions as Metis pull request analysis and a sidecar that monitors indexes, plans and migrations.
AI-written summary · about
chapters (35)
- 0:00intro
- 0:21preface
- 0:47be proactive and push to the left
- 1:58know the context to find the root cause! monitor and troubleshoot
- 2:36how do you know your code will work in production?
- 3:03what can go wrong?
- 6:48ci/cd
- 8:40so what can go wrong?
- 9:06problems with databases
- 11:55slow queries
- 16:55incompatible changes in schema
- 18:55missing indexes
- 19:28too many indexes
- 20:09bugs
- 22:06orm challenges - n plus 1 selects
- 24:18joins
- 24:51polymorphism
- 25:43data types
- 27:56lack of visibility
- 28:54migrations
- 29:51reviews
- 30:44know the context to find the root cause
- 31:59executing the query
- 33:50anatomy of an sql query
- 36:02observability
- 36:26otel
- 37:07traces and spans in otel and jaeger
- 38:56load testing?
- 40:46be proactive and push to the left!
- 41:33metis demo
- 46:13metis features
- 46:36source code integration
- 46:48pull request analysis
- 47:09summary
- 47:55thank you
related talks
54mDatabase Guardrails - new age for developers and databases
22mDatabase Guardrails - New Age for Developers and Databases
37mPlatform Engineers Must Change Developers and Databases - And Here Is How
23mShift-Left: A DevOps and Database Story
50mAdam Furmanek: You Must Test Your Databases And Here Is How
22mData Integrity Tests in CI/CD Pipelines- 23mShift-Left: A DevOps and Database Story
- 23mShift-Left: A DevOps and Database Story