1. What is Site Reliability Guardian for? #
Automatically validate a release with availability, performance and capacity objectives.
The principle #
Site Reliability Guardian (SRG) automates change impact analysis: after a deployment, it checks that the service meets its objectives. A guardian groups objectives (50 at most) around the entities of the service to protect; each validation returns a pass, warning or fail verdict.
2. Key features #
What the app offers.
Overview #
| Feature | Role |
|---|---|
| Objectifs | DQL query and thresholds (fail, warning) on a metric or indicator |
| Seuils auto-adaptatifs | Davis AI computes and adjusts thresholds from history |
| Objectif informatif | Displayed without affecting the verdict |
| Validation insight link | Link to a validation detail, to share in the pipeline |
| Configuration as code | Guardians managed with Monaco or Terraform |
3. Getting started #
First steps.
Step by step #
- Create a guardian on the service to protect.
- Add objectives: error rate, response time, CPU…
- Trigger validation from a workflow at the end of the deployment.
- Block or roll back the release depending on the verdict.
Prerequisites #
- Service data in Grail (metrics, traces, logs).
- The Workflows app to automate validations.
Cloud Automation successor. SRG replaces Cloud Automation quality gates; the documentation describes the migration. Start with informational objectives, then make them blocking once thresholds are reliable.
Sources #
From the official documentation consulted in November 2026: labels and availability change with platform versions.
