Practical guides and frameworks for performance, reliability, and the evidence behind engineering decisions.
LinkedIn carries the short version. The full explanations and diagnostic tools live here.
Interactive demonstrations
A synthetic, adjustable model of queues, worker capacity, and retries. Explore the mechanism without sending traffic to a real system.
Your best marketing campaign can become your worst outage.
Turn up the traffic, watch requests line up, then lower it again. See why the waiting takes longer to clear.
Start with the workflow
Measure the Workflow, Not Just the Request
A blank measurement contract for completion, workload, failures, hidden work, and the decision the result should support.
Practical guides and worksheets
- Before a Performance Fix, Write the Prediction
- Data Growth Is Load: What Request Rate Leaves Out
- Turn “It’s Slow” Into a Performance Investigation Brief
- After the Fix: Choose the Guardrail’s Scope
A reusable decision record for the expected effect, possible regressions, disproof, and next action.
A controlled comparison worksheet for investigating data growth at an unchanged request rate.
An eight-field intake tool, fillable worksheet and illustrative example for choosing the next observation.
A scope comparison and decision record for defining the check, its owner and the evidence needed for reuse.
Measured lab
- Same Query, More Rows: A PostgreSQL Lab
- Repeating the PostgreSQL Lab: What Held Up, What Changed
Generated data · 1,620 recorded queries · charts, raw timings, query plans and reproduction scripts.
A repeated lab shows why growing data and timing variation belong in performance planning, with detailed methods and downloadable evidence.
For measured investigations and results, visit Case Studies.
Core frameworks
- The 7 Silent Killers of System Performance
- The Teardown Stack: My Framework for Debugging Slow Systems
- Why Load Tests Lie (and How to Write One That Doesn’t)
Follow on LinkedIn · Get in touch
When Latency Turns Into LoadRepeating the PostgreSQL Lab: What Held Up, What ChangedSame Query, More Rows: A PostgreSQL LabAfter the Fix: Choose the Guardrail’s ScopeTurn “It’s Slow” Into a Performance Investigation BriefMeasure the Workflow, Not Just the RequestYour best marketing campaign can become your worst outage.