Monitoring
Prometheus Mixins and SLO Burn Alerts
You start from raw counters, graduate to histograms, and finish with SLO documents that link to dashboards humans actually open. The course stresses narrative for incident reviewers, not vanity charts.
Format: Weekend sprint · Timeline: 4 weeks · 30h guided
List price: BRL 1.320 (informational, no checkout on this site)
Caio Rabelo
Release reviewer who still carries a pager for a payments edge stack.
Module map
- Histogram bucket tuning lab with traffic replay
- Recording rule naming conventions that survive grep
- Burn alert pairing matrix for multi-window policies
- Mixin packaging for reuse across clusters
- Annotation discipline for deploy markers
- Silence hygiene workshop with rotation calendars
- Post-incident template tying graphs to customer-visible symptoms
Outcomes we expect to see
Ship a mixin PR consumable by another squad without rewrites
Tune burn alerts that page only when error allocation is threatened
Publish an SLO doc with explicit exclusions your PM accepts
FAQ — includes hard truths
We reference long-term storage patterns but labs stay single-cluster for clarity.
Mentor-reviewed quotes
Burn alert matrix stopped our double paging. Histogram lab felt dense but the replay files helped.
Luana Pires · SRE · EdTech streaming · 5/5 · internal feedback