[Performance]
Performance Signals for Testers
We stay in the tester lane: defining acceptance signals, scripting polite load smokes, and interpreting percentile shifts. Labs use k6 with clear guardrails so you can speak credibly in incident reviews without owning the entire observability stack.
Included focus areas
- Choosing SLIs that map to user journeys
- Scripting smoke tests that will not melt staging
- Dashboard annotations testers can own
- Collaboration cues for SRE hand-offs
- When to stop testing and escalate capacity work
- Ethical load testing policies
- Sample blameless review snippets
Outcomes you can demonstrate
- Author a latency budget proposal for a sample feature
- Run a staged k6 smoke and capture a shareable summary
- List three observability gaps you would flag before major marketing events
Facilitator
Taylor Brooks
Performance coach with SRE experience in retail peaks.
FAQ
Math heavy?
We keep statistics practical; you should be comfortable with averages and percentiles, but calculus is not required.
Cloud costs?
Labs run against provided environments; spinning your own cloud tests is not included in tuition.
What is not covered?
Deep kernel tuning and infrastructure procurement are explicitly out of scope.
Participant notes
Finally understand p95 without feeling like I faked it. k6 scripts are small but intentional—mentor pushed me to delete half my lines.
Good bridge material for QA↔SRE conversations. Would like a bonus on WebSockets, but the workbook links help.