I85PREP
⌘ K
GISI ADO Level study materials
GISI ADO Level study resources

GISI ADO Level: Continuous Delivery and Deployment

GISI ADO Level · Work through your exam, one topic at a time.

All guide sections

Continuous Delivery and Deployment

Scope of this domain

Continuous Delivery and Deployment tests the distinction between always-releasable software and automatic production release, plus the deployment strategies and data-migration rules that make releases uneventful.

Core rules and concepts

  • Continuous Delivery: every change is production-ready and release is a business decision. Continuous Deployment: every passing change goes to production automatically.
  • Pipeline stages: commit, acceptance, staging smoke, production rollout, release. Each promotion uses the same artefact.
  • Strategies: rolling (gradual instance replacement), blue-green (two environments, instant switch and rollback), canary (small traffic slice validated before full exposure).
  • Progressive delivery combines canary traffic shifting, feature flags and automated rollback on SLI breach.
  • Database changes follow expand-contract: add new schema, dual-write, backfill, switch reads, then drop the old structure so old and new versions coexist during rollout.
  • Rollback needs backward-compatible schema and retained artefacts; otherwise roll forward with a fix.
  • Release gates should be automated health checks against error rate and latency, not manual sign-off.

Worked example

A team must rename a column with zero downtime. Reasoning: a direct rename breaks the old version still running during a rolling deploy. Do it in three releases: add the new column with dual writes; backfill and switch reads to the new column; drop the old column once no running version references it. Ship behind a canary at 5% traffic for 30 minutes, comparing error rate and p95 latency to the baseline with automatic rollback so regressions are caught before full exposure.

Traps that cost marks

  • Treating delivery and deployment as identical.
  • Big-bang cutovers with no incremental exposure.
  • Rolling back a destructive migration that already dropped data.
  • Feature flags with no owner or removal date, becoming permanent debt.
  • A canary watched by a human on a dashboard rather than automated analysis.