Skip to main content

Resources and Prompts

The Argo Rollout MCP Server provides MCP Resources for real-time data streams and Prompts for guided workflows.


Available Resources​

Resources provide real-time data streams that can be monitored by AI agents. Read-only context uses resources; state-changing actions use tools.

Resource URIDescription
argorollout://rollouts/listAll rollouts overview across the cluster
argorollout://rollouts/[namespace]/[name]/detailSpecific live rollout status (phase, readyReplicas)
argorollout://experiments/[namespace]/[name]/statusMetrics and progression for live A/B experiments
argorollout://health/summaryCluster-wide orchestration health overview
argorollout://health/[namespace]/[name]/detailsDeep-dive into specific application crash statuses
argorollout://metrics/[namespace]/[service]/summaryPrometheus bounds mapping
argorollout://metrics/prometheus/statusValidation of active metric endpoints
argorollout://history/allGlobal audit trail of all deployed revisions
argorollout://history/[namespace]/[deployment]Specific rollout history trail of old ReplicaSets and images
argorollout://cluster/healthRoot cluster readiness check
argorollout://cluster/namespacesAvailable namespaces map

argorollout://rollouts/list​

Cluster-wide overview of all rollouts.

Use Case:

"Show me an overview of all rollouts in my cluster."
"List all Argo Rollouts across namespaces."
"What rollouts do I have? Fetch the rollouts list."

argorollout://rollouts/[namespace]/[name]/detail​

Specific live rollout status including phase, replicas, conditions, and full YAML manifest.

URI Example: argorollout://rollouts/production/api-service/detail

Use Case:

"What's the current status of the api-service rollout in production?"
"Show me the status of the hello-world rollout in default namespace."
"Is my frontend rollout healthy? Check the status in staging."

argorollout://experiments/[namespace]/[name]/status​

Metrics and progression for live A/B experiments.

URI Example: argorollout://experiments/production/api-ab-test/status

Use Case:

"What is the status of the api-ab-test experiment in production?"
"Show me the analysis results for the ui-experiment experiment in staging."

argorollout://health/summary​

Cluster-wide orchestration health overview.

Use Case:

"Give me a health summary of the entire cluster."

argorollout://health/[namespace]/[name]/details​

Deep-dive into specific application crash statuses and health score.

Use Case:

"Show health details for api-service in production."

argorollout://metrics/[namespace]/[service]/summary​

Prometheus bounds mapping for request rate, error rate, latency.

Use Case:

"Show performance metrics for api-service in production."

argorollout://metrics/prometheus/status​

Validation of active metric endpoints.

Use Case:

"Is Prometheus connected and working for metrics collection?"

argorollout://history/[namespace]/[deployment]​

Specific rollout history trail of old ReplicaSets and images.

Use Case:

"Show me the deployment revision history for api-service rollout in production including previous images."
"What are the last rollout history entries and replica set hashes for frontend in staging?"

argorollout://history/all​

Global audit trail of all deployed revisions.

Use Case:

"Show me all deployment history across the cluster."

argorollout://cluster/health​

Root cluster readiness check.

Use Case:

"What is the overall health and capacity of my Kubernetes cluster?"

argorollout://cluster/namespaces​

Available namespaces map.

Use Case:

"List all namespaces in the cluster."

Available Prompts​

Prompts are guided workflows that orchestrate multiple tools to accomplish complex tasks.

PromptDescription
onboard_application_guidedZero YAML first-time conversion (Deployment → Rollout; direct or workloadRef)
canary_deployment_guidedProgressive traffic execution playbook (5% → 100%)
blue_green_deployment_guidedZero-downtime cutover workflows with instant rollback hooks
rolling_update_guidedSafe traditional rollout patterns
cost_optimized_deployment_guidedCost-conscious best practices (orchestration tools are future enhancement)
multi_cluster_canary_guidedSequential regional rollout; switch kubeconfig context per region

onboard_application_guided​

Zero YAML first-time conversion from Deployment to Rollout. Full Step 1–5 walkthrough: validate, convert, link traffic (optional), setup analysis (optional), verify.

Use Case:

"Onboard api-service in production to Argo Rollouts"
"Migrate my api-service deployment in production to Argo Rollouts safely."

canary_deployment_guided​

Progressive canary deployment playbook. Pre-flight → deploy → weight shifts (5% → 10% → 25% → 50% → 100%) → promote/abort.

Use Case:

"Guide me through a canary deployment of api-service:v2 in production"
"Deploy api-service:v2 with canary strategy — 5% then 100%."

blue_green_deployment_guided​

Blue-green cutover with instant rollback. Preview → analysis → switch traffic.

Use Case:

"Set up a blue-green deployment for checkout in production"
"Convert the frontend in staging to a blue-green rollout. Auto-promote after 300s."

rolling_update_guided​

Standard rolling update patterns. Validate → update → monitor.

Use Case:

"Do a rolling update of frontend in staging to image frontend:2.0"
"Deploy hello-world:v2 to the hello-world rollout in default."

cost_optimized_deployment_guided​

Cost-conscious best practices. Orchestration tools are a future enhancement.

Use Case:

"Deploy api-service:v2 in production with cost optimization"

multi_cluster_canary_guided​

Sequential regional rollout. Switch kubeconfig context per region.

Use Case:

"Run an intelligent multi-cluster canary for api-service:v2 across regions"

Using Resources and Prompts​

Accessing Resources​

Resources are accessed via the MCP protocol using their URI. Read-only context (list rollouts, status, health, metrics) uses resources.

Client: Show me the status of the api-service rollout in production
Agent: [Subscribes to argorollout://rollouts/production/api-service/detail]
[Returns rollout phase, replicas, conditions]

Using Prompts​

Prompts are invoked to guide complex multi-step operations. State-changing actions (create, update, delete, promote, abort) use tools.

Client: Migrate my api-service deployment in production to Argo Rollouts
Agent: [Invokes onboard_application_guided prompt]
[Orchestrates validate → convert → link traffic → analysis → verify]
[Provides step-by-step progress updates]

Next Steps​