Canary deployment

As explained in Infrastructure as Code, Patterns and Practices: With examples in Python and Terraform

A canary deployment is a release pattern that incrementally exposes a change to a limited portion of users, traffic, or infrastructure before expanding it broadly.

A canary deployment sends a small percentage of traffic to updated resources before expanding exposure. Successful requests justify increasing the percentage; errors trigger debugging and a return to the previous environment. section 9.2.3

Apply the pattern as a controlled rollout:

• Prepare updated green resources alongside the existing blue resources. section 9.2.3
• Route a limited share of requests to green, such as 10% green and 90% blue. section 9.2.3
• Observe errors and functionality, then increase green traffic gradually when results remain acceptable. section 9.2.4
• Reduce green traffic and debug when failures appear; retain blue until the regression window provides confidence. section 9.2.4

Canarying a database change

For stateful infrastructure, applications first write to both blue and green databases, then write exclusively to green, and finally read from green. Automation may reconcile duplicated data, while incremental rollout preserves a route back to blue. section 9.3.3

Figure 9.7 Configure the global load balancer to run a canary test and send a small percentage of traffic to green resources.

Shows a global load balancer sending a small traffic percentage to green resources during a canary test. section 9.2.3

How Infrastructure as Code, Patterns and Practices: With examples in Python and Terraform covers it

How deep it goes

Explains traffic weighting, rollback, monitoring, regression windows, and stateful cutover, but leaves increment selection context-dependent.

Examples

Uses blue-green networks, weighted load-balancer services, application instances, and staged database writes and reads.

What it assumes

Assumes parallel blue and green resources plus a mechanism for directing traffic between them.

What you can do after

Operate/configure incremental traffic shifts, monitor regressions, and revert to the previous environment when failures emerge.

How it connects

Connects canary deployment with blue-green deployment, regression testing, load balancing, infrastructure as code, and database migration.

liveBookmanning.com

Read Infrastructure as Code, Patterns and Practices: With examples in Python and Terraform on liveBook Get the book at manning.com

Also explained in

• 

  liveBookmanning.com
  Platform Engineering on KubernetesA canary release is a deployment technique that gradually routes live traffic to a new software version while retaining traffic on a stable version for early validation and rollback.
• 

  liveBookmanning.com
  Kubernetes in Action, Second EditionA release technique that exposes a new application version to a limited subset of instances or traffic before expanding it broadly.
• 

  liveBookmanning.com
  Machine Learning System DesignCanary deployment is a release strategy that sends a limited portion of users or traffic to a new version before fully switching over.
• 

  liveBookmanning.com
  Google Anthos in ActionA canary deployment is a release strategy that exposes a new version to a small portion of traffic before expanding its availability.
• 

  liveBookmanning.com
  Think Like a Software Engineering ManagerA canary deployment releases a software change to a limited subset of users or infrastructure before expanding it to the full population.
• 

  liveBookmanning.com
  Kubernetes for DevelopersA rollout is the controlled propagation of an application or platform update across its deployed instances or locations.

Compare how 7 books explain canary deployment →

Related concepts

• blue-green deploymentalternative to · Blue-green deployment is a release pattern that maintains separate active and replacement environments and switches traffic between them after the replacement is validated.
• deploymentkind of · Deployment is the process of making software, models, or other computational systems available to operate in a target environment.
• testing in productionused by · Testing in production is the practice of evaluating software or infrastructure behavior in a live production environment.
• evaluation suiteuses · An evaluation suite is an organized collection of tests, cases, or tasks used to assess a system against specified behaviors or criteria.
• experimentationcontrasts with · Experimentation is the deliberate testing of an idea, intervention, or alternative by observing outcomes to gather evidence or compare effects.

This explainer was generated by Manning's knowledge graph from the text of Infrastructure as Code, Patterns and Practices: With examples in Python and Terraform. The section links point to the parts of the book it draws on.