How-To
Alert trust: decide which signals earn an interruption
Decide which alerts earn an interruption using urgency, reachable authority, coverage evidence, and a practical alert acceptance record.
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
Canary deployments: limit exposure and test the comparison
Choose representative traffic, define promotion criteria, and verify rollback compatibility before expanding a release.
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.
AIOps
Python logging for operations: record failure and preserve the exit status
Use Python logging to capture useful context without swallowing errors, duplicating exceptions, or exposing sensitive data.
AIOps
Collect and aggregate operational data with Python
Validate the source data, preserve useful denominators, and distinguish a failed collection from a real zero before aggregating metrics.
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
Distributed tracing: choose the first request path to instrument
Start tracing with a customer-critical path, then test propagation, sampling, and whether the trace supports a real investigation.
AIOps
Route support tickets with NLP and a reliable fallback
Build a ticket-routing evaluation around real ownership, uncertain cases, and correction cost. Start with an honest baseline before adding a model.