Examples and Quick Reference
Quick reference for natural language prompts, workflow triggers, and command patterns. For detailed step-by-step guides, see Common Workflows.
Example Requests
Backend Discovery
"What Prometheus backends are available?"
"Show me the health status of our production Prometheus."
"What features are enabled on our Thanos instance?"
PromQL Querying
"What is the current request rate for http_requests_total?"
"Show me CPU usage across all pods in production over the last hour."
"Calculate p99 latency from the http_request_duration_seconds histogram."
"What metrics does the api-server expose?"
Application Onboarding
"What's the best way to monitor my Python Flask app on Kubernetes?"
"Wire my api-server service in the default namespace to Prometheus."
"Validate the /metrics endpoint for my app at http://my-app.default:8080/metrics."
"Add my VM host 10.0.1.5:9100 to Prometheus file_sd targets."
Exporter Deployment
"Deploy a postgres_exporter in the monitoring namespace for our PostgreSQL database."
"Install node_exporter as a DaemonSet across all nodes."
"What exporters are available for Redis?"
"Verify the postgres_exporter is healthy and being scraped."
Rule Management
"Draft an alert for when the 5xx error rate exceeds 5% for 5 minutes."
"Validate this alerting rule YAML before deploying it."
"If I set the threshold to 3%, would this alert have fired in the last 7 days?"
"Find the CRD that owns the alertmanager.rules group and update the threshold."
TSDB FinOps
"How many active series does our Prometheus have?"
"Analyze cardinality hotspots and recommend optimization strategies."
"Generate a relabel config to drop the pod_id label from http_requests_total."
"Configure remote-write to our Thanos receive endpoint."
Troubleshooting
"Why is my api-server job showing as down in Prometheus?"
"Which scrape targets are currently failing?"
"Show me all targets with connection refused errors."
Synthetic Monitoring
"Monitor https://talkops.ai availability from inside the cluster."
"Set up a Blackbox exporter probe for our health check endpoint."
Command Reference Guide
| User Intent | Core Tools | Example Query |
|---|---|---|
| Discover backends | prom://system/backends | "What Prometheus backends are available?" |
| Query metrics | prom_query_instant / prom_query_range | "Show request rate for the last hour" |
| Explore labels | prom_explore_labels | "What labels does http_requests_total have?" |
| Validate query | prom_validate_promql | "Check if this PromQL is valid" |
| Onboard app | prom_recommend_instrumentation → prom_test_endpoint → prom_apply_servicemonitor | "Set up monitoring for my Python app" |
| Deploy exporter | prom_install_exporter → prom_verify_exporter | "Deploy postgres_exporter to monitoring namespace" |
| Draft alert | prom_draft_alert_rule → prom_check_rule_group | "Alert on high error rates" |
| Simulate alert | prom_simulate_firing_historical | "Would this have fired last week?" |
| Patch CRD rule | prom://kubernetes/prometheusrules → prom_upsert_rule_group | "Update the alertmanager rules threshold" |
| Analyze cardinality | prom_optimize_cardinality → prom_plan_relabel | "Which metrics cost the most?" |
| Fix failed target | prom://topology/failed_targets → prom_test_endpoint | "Why is my job down?" |
| Configure remote-write | prom_configure_remote_write | "Send metrics to Thanos" |
| Monitor endpoint | prom_install_exporter → prom_apply_probe | "Monitor https://example.com availability" |
Tips and Best Practices
Start with Resources, Then Use Tools
Always read resources for context before making tool calls:
"Show me all available backends" → prom://system/backends
"What metrics does the api-server expose?" → prom://topology/services/api-server/metrics
Validate Before Executing
For complex PromQL, validate syntax first:
"Validate this query: sum(rate(http_requests_total{status=~'5..'}[5m])) / sum(rate(http_requests_total[5m]))"
Use Guided Prompts for Complex Workflows
The server includes 5 built-in guided prompts for multi-step workflows:
| Prompt | Use Case |
|---|---|
prom-k8s-app-onboarding-guided | Onboard a K8s app from scratch |
prom-k8s-exporter-onboarding-guided | Deploy and wire an exporter |
prom-vm-legacy-onboarding-guided | VM/legacy target setup |
prom-query-guided | Safe metric querying |
prom-troubleshoot-guided | Diagnose failed targets |
Cross-Reference CRDs Before Upsert
When patching PrometheusRule CRDs, always check prom://kubernetes/prometheusrules first to get the exact name and namespace:
"Find the CRD that owns the alertmanager.rules group."
→ Then: "Update the rule threshold in that CRD."
Next Steps
- Common Workflows — Step-by-step workflow guides
- Tools — Available MCP tools reference
- Resources — MCP resources for monitoring
- Configuration — Server configuration