autopilot-deploy

Automate four-stage canary deployments with SLO monitoring and circuit breakers.

18|4|Updated Apr 29, 2026

Search Tags:#deployment #Kubernetes #canary release #progressive rollout #SLO monitoring

Ready to run

npx skills add https://github.com/aws-samples/sample-oh-my-aidlcops --skill autopilot-deploy

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: autopilot-deploy
Source: https://github.com/aws-samples/sample-oh-my-aidlcops/tree/main/plugins/agenticops/skills/autopilot-deploy
Command: npx skills add https://github.com/aws-samples/sample-oh-my-aidlcops --skill autopilot-deploy

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires argoproj/argo-rollouts, flagger, prometheus, and includes scripts (resource) and references (resource) components.

What problem does it solve?

This Skill automates the progressive deployment of agents or skills to production environments, ensuring a smooth transition with minimal risk and manual intervention.

Core Features & Use Cases

4-Stage Progressive Deployment: Automates the deployment process in four stages (1% → 10% → 50% → 100%) with continuous evaluation and circuit breaker mechanisms.
Canary Release Strategy: Implements a canary release strategy for new agent versions or skill updates, reducing risk by gradually increasing traffic to the new version.
Continuous Evaluation: Utilizes continuous evaluation checks to ensure quality and performance before each deployment stage.
Circuit Breaker: Automatically rolls back to the previous stage if any SLO is violated, ensuring the stability of the production environment.

Quick Start

Deploy a new version of an agent or skill to production using the autopilot-deploy skill with the target deployment target, e.g., /autopilot-deploy rag-qa-agent:v2.3.1

Frequently Asked Questions about autopilot-deploy

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I automate progressive canary deployments in Kubernetes?▼

You can automate progressive canary deployments in Kubernetes using a four-stage rollout strategy that shifts traffic from 1% to 100% with continuous SLO evaluation and circuit breakers for automatic rollbacks.

What is a canary release strategy with continuous SLO monitoring?▼

A canary release strategy with continuous SLO monitoring gradually routes production traffic to new agent versions while tracking metrics via Prometheus, automatically rolling back if service level objectives are violated.

Do I need Argo Rollouts or Flagger to automate production deployment progression?▼

Yes, automating progressive production deployment progression in Kubernetes requires either Argo Rollouts or Flagger to manage the four-stage traffic shifting and circuit breaker mechanisms.

How does a circuit breaker handle rollback during a Kubernetes canary release?▼

A circuit breaker handles rollback during a canary release by automatically reverting traffic to the previous stable stage whenever Prometheus SLO monitoring detects performance degradation or quality violations.

Can I use Prometheus to monitor deployment stages and trigger rollbacks?▼

Yes, you can use Prometheus to monitor deployment stages by running continuous SLO checks at each phase, triggering a circuit breaker rollback if metrics fail before reaching the 100% traffic allocation.