Run / Reliability
Production Reliability
Finding what is actually breaking in production, then building the observability to see it coming next time.
Most reliability work starts with a disagreement about what is wrong. Support thinks it is the database, engineering thinks it is traffic, and nobody has the evidence to settle it. The first job is measurement, not fixes.
From there it is unglamorous: thread and heap dumps read properly, JVM tuning, cluster sizing that matches real traffic rather than a guess, and release coordination so deployments stop being the most dangerous hour of the week.
I currently lead L2 production support for a B2B storefront that cannot go down. That includes carrying the consequences of architectural decisions — which is why I make different ones.
Where this has been done
Application Support Lead
Lead the L2 production support team. Thread and heap dump analysis, AppDynamics and Splunk, release coordination, and a data centre migration with minimal disruption.
Technical Architect
Kept the site up through traffic growth — JBoss tuning, cluster expansion, and analysing what actually fell over.