How-To
Canary deployments: limit exposure and test the comparison
Choose representative traffic, define promotion criteria, and verify rollback compatibility before expanding a release.
Guides
Walkthroughs for observability, incident response, and AI-assisted operations.
AIOps fundamentals
Start with what AI can do for operations, and where judgment still matters.
Read AIOps fundamentals ↗Observability for SRE
Make metrics, logs, and traces work together.
Read Observability for SRE ↗Incident management with AI
Build an incident response practice that learns.
Read Incident management with AI ↗Practical articles
Instructions, examples, and implementation notes.
How-To
Mean time to detect: calculate MTTD and expose measurement gaps
Use consistent onset and discovery timestamps, include customer-reported incidents, and interpret the average with its sample and uncertainty.
How-To
Metrics for observability: choose the question before the chart
Counters, gauges, and distributions answer different questions. Build metrics that preserve scope, denominators, and the evidence behind an incident.
How-To
Observability logs: preserve the events an investigation needs
Design logs around useful events, searchable context, and a collection path whose failures are visible.
How-To
Design an on-call rotation people can actually support
Define coverage, escalation, training, and recovery time before assigning the calendar. Review workload as well as shift counts.
SRE
SRE runbook template: diagnosis, bounded action, and recovery
Download a Markdown runbook template and follow a database connection example that separates symptoms, mitigation decisions, and verified recovery.
How-To
Containers and orchestration: understand what the platform guarantees
Containers package processes; orchestration reconciles desired state. Reliability still depends on probes, capacity, dependencies, and application behavior.
How-To
Read a distributed trace without mistaking it for the whole system
Understand spans, parent-child relationships, retries, and missing evidence so traces support a careful diagnosis.
How-To
Read Google’s free SRE books with an operational question in mind
Use the free SRE books as references for a concrete service problem, then adapt and test the practices against your own constraints.