Team collaborating around laptops

How a profiling week unfolds

This page mirrors the checklist we send after you sign a statement of work. It is the same sequence whether we work remote or hybrid from Cape Town.

  1. Access and safety briefing

    VPN, read-only tokens, and written rules for production touch. We confirm rollback contacts and quiet hours.

  2. Symptom charter

    One paragraph metric definition: which endpoint, which queue, which batch window fails. Everyone signs off so we do not chase the wrong graph.

  3. Reproduction harness

    Fixtures, recorded messages, or scripted load. Goal is a repeatable run that fails within 30 minutes on a laptop or staging pod.

  4. Baseline capture

    CPU and memory profiles stored in your repo under docs/perf/ or a path you choose. We never keep exclusive copies.

  5. Pair reading sessions

    Two live calls where we walk through cumulative time and allocation stacks with your developers. Questions welcome; silence is fine too.

  6. Patch proposals

    Small PRs with before/after profile snippets in the description. Large rewrites only when data supports them.

  7. Verification

    Re-run the harness under agreed concurrency. If metrics miss target, we document what remains and optional follow-up scope.

  8. Handoff

    Recorded walkthrough plus hotspot memo PDF for auditors. Optional on-site whiteboard session at 14 Adderley Street.

Ready to reserve a week?

Send your symptom charter draft—even bullet points—and we respond with open dates and a fixed quote.

Submit a charter draft Compare engagement types