intermediate · self-paced
Indexing Internals Workshop
4 weeks · 6–9 hrs/week
Move from “add an index and hope” to decisions you can justify. Labs use synthetic datasets with skew and contention so you can see planner behavior without risking production.
What ships in this outline
- Hands-on EXPLAIN workshops with before/after captures
- Partial and covering indexes with measurable wins
- Vacuum/bloat awareness at a practitioner level
- Workload cards describing read vs write sensitivity
- Checklist for index review in pull requests
- Office hours for slow-query triage practice
Outcomes
- Pick index types for three contrasting access patterns
- Present a one-page rationale a staff engineer would accept
- Spot redundant indexes from metadata alone
Lead mentor
Lin Okada
SRE turned database mentor; previously owned incident reviews for query regressions.
FAQ
PostgreSQL is primary; concepts map to MySQL with callouts where divergences matter.
No. You practice on provided clusters; production tuning stays in mentorship calls, not graded labs.
Columnstore and specialized OLAP engines are out of scope for this workshop.
Experience notes
“The skewed dataset lab mirrored a bug we had in staging. Seeing the planner flip was the aha moment.”