Examples & Workflows
Things you can ask the assistant to do, and how it gets there.
Installing a chart​
You say: "Install PostgreSQL from Bitnami in the database namespace"
What happens: The assistant runs search → get chart info → validate values → dry-run → install → monitor. It follows the safe deployment workflow automatically.
| Step | Tool |
|---|---|
| 1 | helm_search_charts — find the chart |
| 2 | helm_get_chart_info — get metadata and values |
| 3 | helm_validate_values — validate config |
| 4 | helm_dry_run_install — preview |
| 5 | helm_install_chart — install |
| 6 | helm_monitor_deployment — watch rollout |
Upgrading a release​
You say: "Upgrade my-app to version 2.0 with 3 replicas"
What happens: Check status → get new version info → validate → upgrade → monitor.
Rolling back​
You say: "Rollback my-release to the previous version"
What happens: Get release history → pick target revision → rollback → verify.
Troubleshooting​
You say: "My pods are in CrashLoopBackOff after deploying redis"
What happens: The assistant pulls helm_troubleshooting_guide with error_type="pod-crashloop" and walks through the diagnosis.
More things you can ask​
Discovery: "Search for nginx ingress charts", "Get info about prometheus-stack", "Show me the README for bitnami/postgresql"
Installation: "Install redis in the cache namespace with 3 replicas", "Dry-run installing ArgoCD to see what resources will be created"
Status & monitoring: "List all Helm releases in production", "What's the status of my-redis?", "Monitor the nginx deployment"
Cluster ops: "What Kubernetes contexts are available?", "Switch to the production cluster", "List all namespaces"
Validation: "Validate my values for the PostgreSQL chart", "Check dependencies for kube-prometheus-stack", "Show me the installation plan for ArgoCD"
Guidance: "What are the security best practices for Helm?", "How do I troubleshoot image pull errors?", "Guide me through upgrading my PostgreSQL release"
End-to-end workflows​
Production deploy with validation: Search → get install plan → validate values → dry-run → install → monitor → verify.
Upgrade with rollback plan: Get current status → review upgrade changes → note revision for rollback → upgrade → verify. If it breaks: rollback.
Multi-cluster: List contexts → switch to staging → install → verify → switch to prod → install with prod values → verify.
Security-first: Show security checklist → validate chart source → validate values → render manifests for review → install in dedicated namespace.
Tips​
- Dry-run first — "Dry-run the installation first to see what will be created"
- Pin versions — "Install bitnami/postgresql version 12.5.0"
- Use namespaces — "Install redis in the cache namespace"
- Validate before deploy — "Validate my values before installing"
- Monitor after deploy — "Monitor the deployment until healthy"
Next steps​
- Tools — MCP tools reference
- Resources & Prompts — MCP resources and prompts
- Configuration — Server configuration