Available Tools
Complete reference for all MCP tools provided by the Argo Rollout MCP Server. Tools are organized by category.
Migration & Generations​
Tools for converting Deployments to Rollouts and managing migration workflows.
| Tool | Description |
|---|---|
convert_deployment_to_rollout | Convert K8s Deployment → Argo Rollout (direct or workloadRef migration; auto-applies Services and CRDs) |
convert_rollout_to_deployment | Reverse migration back to standard K8s Deployments |
argo_manage_legacy_deployment | Unified: scale, delete, or generate scale-down manifest for legacy Deployments (workloadRef migration) |
create_stable_canary_services | (Advanced/legacy) Generate stable+canary Services. Prefer convert_deployment_to_rollout(mode='generate_services_only', app_name='...') |
generate_argocd_ignore_differences | Create ArgoCD sync rules for safe Rollout integrations |
validate_deployment_ready | Check Deployment readiness score before touching the cluster |
validate_deployment_ready​
Unified pre-flight check: validates Deployment structure (selector, template, containers, replicas, limits, probes) and Service selector compatibility (no pod-template-hash). Single call for full migration readiness.
"Check if my hello-world deployment in the default namespace is ready to migrate to Argo Rollouts."
"Run a readiness check on deployment api-service in namespace production before converting it to a rollout."
"Validate the frontend deployment in staging — does it meet all requirements for Argo Rollouts onboarding?"
convert_deployment_to_rollout​
The correct tool for onboarding an existing app. Auto-fetches the Deployment from the cluster, preserves all config (resource limits, probes, env vars), auto-discovers the existing Service's port and selector, and with apply=True creates the Rollout CRD + stable/canary Services in a single call.
"Convert the api-service deployment in production to a canary Argo Rollout and apply it to the cluster."
"Migrate hello-world in default to an Argo Rollout using canary strategy — apply everything to the cluster now."
"Convert api-service in production to a canary rollout using workloadRef mode (no pod duplication), scale down on success, apply=True."
"Convert api-service in production to a canary rollout — give me the YAML for review first, don't apply yet."
For brand-new apps (no existing Deployment): use argo_create_rollout instead.
convert_rollout_to_deployment​
Reverse migration back to standard Kubernetes Deployments.
"Convert the api-service Argo Rollout in production back to a standard Kubernetes Deployment with RollingUpdate strategy, 25% max surge."
"I need to abandon Argo Rollouts for frontend — convert the rollout YAML back to a standard deployment."
generate_argocd_ignore_differences​
Create ArgoCD ignoreDifferences configuration for safe Rollout integrations.
"Generate the ArgoCD ignoreDifferences configuration for api-service in production — include Rollout status and AnalysisRun fields."
"Generate ignoreDifferences for api-service in production with include_deployment_replicas for workloadRef."
"Create the ArgoCD ignoreDifferences snippet for checkout in production — include Rollout status, AnalysisRun, and optionally TraefikService."
argo_manage_legacy_deployment​
Unified tool for Deployment lifecycle during workloadRef migration. Use action: scale_cluster, delete_cluster, or generate_scale_down_manifest.
"Generate a scale-down manifest for the api-service deployment in production — I'll commit it to Git for Argo CD to apply."
"Scale the api-service deployment in production to 0 replicas."
"Delete the api-service deployment in production — we've fully migrated to the Rollout."
Lifecycle Orchestration & Operations​
Tools for managing Rollout lifecycle: create, update, promote, pause, resume, abort.
| Tool | Description |
|---|---|
argo_create_rollout | Create new Argo Rollout (canary, bluegreen, rolling) |
argo_delete_rollout | Safely remove a Rollout from the cluster |
argo_update_rollout | Unified: update image (direct or workloadRef), strategy, traffic routing, or workloadRef scale-down |
argo_manage_rollout_lifecycle | Unified lifecycle: promote, promote_full, pause, resume, abort, skip_analysis |
argo_create_rollout​
Create new Argo Rollout with canary, blue-green, or rolling strategy.
"Create a new canary Argo Rollout payment-service in production with image payment:v1 on port 8080."
argo_update_rollout​
Unified tool for updating image, strategy, traffic routing, or workloadRef scale-down. Use update_type: image, strategy, traffic_routing, workload_ref.
Image update:
"Deploy api-service:v2.0 to the api-service rollout in production."
"Update the frontend rollout in staging to image frontend:2.1.0."
"Roll out a new version of checkout — new image is checkout:1.6-hotfix in production."
Traffic routing (canary only):
"Link the hello-world rollout in default to TraefikService hello-service-route-wrr."
"Set trafficRouting on api-service rollout in production to use TraefikService api-service-route-wrr."
Strategy update:
"Update the canary-demo rollout in canary-demo namespace to use canaryService canary-demo-preview and steps: 20% → pause → 40% → pause 10s → 60% → pause 10s → 80% → pause 10s."
argo_manage_rollout_lifecycle​
Unified tool for rollout lifecycle actions. Use action: promote, promote_full, pause, resume, abort, skip_analysis.
"Promote api-service rollout in production to the next step."
"Skip all remaining steps and promote api-service rollout fully to 100%."
"Pause the api-service rollout in production — hold at current traffic level."
"Resume the paused api-service rollout in production."
"Abort the api-service rollout in production and roll back to stable."
"Something is wrong with the canary — roll back frontend rollout in staging immediately."
"Emergency: skip the analysis step and promote api-service rollout in production — Prometheus is down but the version is verified healthy."
argo_delete_rollout​
Safely remove a Rollout from the cluster.
"Delete the api-service Argo Rollout in production — we're reverting to standard Deployments."
"Remove the frontend rollout in staging and all its associated services."
Validation, Traffic & Observation​
Tools for configuring analysis, A/B experiments, and validation.
| Tool | Description |
|---|---|
argo_configure_analysis_template | Configure AnalysisTemplate: execute (create+link) or generate_yaml (GitOps review) |
argo_create_experiment | Instantiate ephemeral A/B test pods |
argo_delete_experiment | Clean up experimentation runs |
argo_configure_analysis_template​
Configure Prometheus-based AnalysisTemplate. Use mode: execute (create+link) or generate_yaml (GitOps review).
"Set up automated Prometheus analysis for the api-service rollout in production with Prometheus at http://prometheus:9090. Abort if error rate exceeds 5% or P99 latency exceeds 2 seconds."
"Configure analysis for my hello-world rollout using Prometheus at http://prometheus.monitoring:9090."
"Add automated health checks to the api-service rollout — link it to Prometheus so failing canaries auto-rollback."
argo_create_experiment​
Create an A/B test experiment with baseline and candidate templates.
"Create an A/B test experiment called api-ab-test in production — run baseline (stable) and candidate (canary) side by side for 30 minutes."
"Start an Argo Experiment named ui-experiment in staging with two templates: control (stable spec) and variant (canary spec), running for 1 hour."
argo_delete_experiment​
Clean up experimentation runs.
"Delete the api-ab-test experiment in production — the baseline won, keeping stable."
"Clean up the ui-experiment in staging."
Tool Availability Summary​
| Category | Total Tools |
|---|---|
| Migration & Generations | 6 |
| Lifecycle Orchestration & Operations | 4 |
| Validation, Traffic & Observation | 3 |
| Total | 13 |
Next Steps​
- Resources and Prompts — MCP resources and prompts
- Common Workflows — Step-by-step workflow guides
- Examples — Quick reference and prompts