How I work with it
Metrics-driven operations: instrument services, record SLIs, alert on symptoms not causes, and keep cardinality honest so Prometheus stays fast and bills stay boring.
Knowledge areas
- PromQL for real SLO math — error budgets, burn rates
- Exporters, service discovery and federation
- Alert design that respects on-call sanity
knowledge shown honestly — no percentages, no infographics.
want the deep-dive version? ask me