Skip to main content

Module 4: Operational Readiness & Security Review: Worked Examples

Example 1: Define an SLO From User Harm​

Measure successful critical operations and latency below a justified threshold. Set a target and budget, exclude only explicitly non-user-impacting traffic, and link alerts to burn and action. CPU is diagnostic context, not the user outcome.

Example 2: Rehearse a Bad Deploy​

Inject an error regression, detect it through the intended alert, follow the runbook, decide rollback, handle schema compatibility, communicate, and verify recovery. Update the runbook where the operator hesitated.

Example 3: Prove Recovery, Not Backup Creation​

Choose RPO/RTO before mechanism. Restore into an isolated environment, validate schema and representative records, time the process, and record gaps. An untested backup is only a hope.

Completion Standard​

  • Produce SLO/dashboard/alert evidence, threat model, least-privilege review, exercised runbook, and timed restore proof.