liveBook · Manning

Analysis template

As explained in Platform Engineering on Kubernetes

An analysis template is a reusable Argo Rollouts specification that defines automated measurements and evaluation criteria for validating a release.

An AnalysisTemplate is a reusable Argo Rollouts specification for automated measurements and evaluation criteria. It defines probes, providers, success conditions, and failure limits that gate release progression. section 8.3.3

Use an AnalysisTemplate to control when an AnalysisRun is created and how its result affects a rollout:

  • Query service metrics through providers such as Prometheus, Datadog, New Relic, or Dynatrace. section 8.3.3

Reusable success-rate gate

Define a success-rate AnalysisTemplate with a service-name argument, a five-minute interval, a Prometheus provider, the condition result[0] >= 0.95, and failureLimit: 3. Reference it from a canary rollout and supply the canary service address as the argument.

How Platform Engineering on Kubernetes covers it

How deep it goes
Defines providers, queries, intervals, arguments, success conditions, and failure limits, but not provider internals.
Examples
Uses a Prometheus success-rate template, canary starting steps, explicit analysis steps, and blue-green smoke tests.
What it assumes
Assumes Argo Rollouts, rollout strategies, Kubernetes services, and an available metrics provider.
What you can do after
Configure it to automate release checks and prevent traffic progression when measured behavior violates defined conditions.
How it connects
Connects Argo Rollouts with AnalysisRuns, canary releases, blue-green deployments, Prometheus, and service metrics.

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Platform Engineering on Kubernetes. The section links point to the parts of the book it draws on.