Step 01 of 06
Learn the concept
A deployment that depends on a person remembering fourteen commands is not a process. It is folklore with a pager attached. A pipeline turns the path from commit to running service into evidence: what changed, what checked it, what artifact moved, and how quickly recovery happened.
The ideas this is made of
CI answers whether the change is safe to merge
Continuous integration is about integrating code with other code. It runs on pushes and pull requests before the change is trusted. A good CI run is boring: format, vet, tests, race detector, build. It says the proposed source is coherent enough to become a candidate.
Delivery prepares a releasable artifact
Continuous delivery means every accepted change can become a release because the artifact is built, scanned, signed and stored. A human or policy may still approve production. If release day needs a special rebuild or secret command, the system is not continuously deliverable.
Deployment changes what users run
Continuous deployment is the stronger promise: a passing change reaches users automatically. That can be excellent for small, observable services and reckless for systems with manual data migrations. During an incident, 'delivered' and 'deployed' are not trivia.
DORA metrics keep the argument honest
Lead time and deployment frequency measure flow. Change failure rate and time to restore measure safety. Optimizing only speed creates a conveyor belt into incidents. Optimizing only safety creates a museum. The four together show whether delivery improved.
lead time: order placed -> drink on counter
deployment frequency: batches served per hour
change failure rate: drinks remade
time to restore: minutes until queue recoversThe same four measurements work outside software. A shop can be fast and wrong, slow and safe, or fast with quick recovery. Measurement makes the machinery useful.
CI, delivery and deployment are not synonyms
| Practice | Trigger | Ends when | Risk moved |
|---|---|---|---|
Continuous integration | Push or PR | Checks pass | Bad code is stopped early |
Continuous delivery | Merge or release request | Artifact is releasable | Release is a decision |
Continuous deployment | Passing gate | Users run it | Rollback and observability matter |
Manual deploy | Someone remembers | Commands finish | Human memory carries risk |
What these are called on the job
Trust boundary — The line between source someone wrote and an artifact allowed to affect users.
Lead time — Elapsed time from committed change to running in an environment.
Change failure rate — The share of deployments that cause incidents, rollbacks or urgent fixes.
Time to restore — How long the team needs to return service after a bad change.