Workflow: Application Onboarding

Step-by-step guide for onboarding applications to Prometheus monitoring — from choosing the right instrumentation strategy through endpoint validation, ServiceMonitor creation, and data verification.

---

When to Use

Use this workflow to:

• Instrument a custom application (Go, Java, Python, Node.js) for Prometheus
• Deploy an exporter for a third-party system (PostgreSQL, Redis, NGINX)
• Add VM/legacy targets via file_sd_configs
• Set up synthetic endpoint monitoring with Probe CRDs

---

Workflow 1: Kubernetes App Onboarding

Journey

Step-by-Step

Step	Action	Tool / Resource	Key Parameters
1	Verify backend health	Resource: prom://system/backends/{backend_id}	Confirms connectivity, features, retention
2	Choose instrumentation strategy	prom_recommend_instrumentation(workload_type="custom_app", language="python", environment="kubernetes")	Returns: direct_instrumentation / exporter / builtin_metrics
3	Deploy instrumented app	(Manual) — add instrumentation, build, deploy to K8s	—
4	Validate /metrics endpoint	prom_test_endpoint(endpoint_url="http://my-app.default:8080/metrics")	Returns: ok, metrics_count, format, errors
5	Apply ServiceMonitor	prom_apply_servicemonitor(namespace="default", service_name="my-app")	Auto-detects Prometheus Operator selector labels
6	Discover new metrics	prom_explore_labels(backend_id="default", metric_name="http_requests_total")	Shows all label names and top values
7	Verify data flows	prom_query_instant(backend_id="default", query="rate(http_requests_total[5m])")	Confirms metrics are being ingested

Guided Prompt: Use prom-k8s-app-onboarding-guided for the full step-by-step flow with the AI.

Strategy Decision Matrix

Workload	Language	Framework	Recommended Strategy
Custom app	Go/Java/Python/Node	—	direct_instrumentation
Custom app	Java	Spring Boot	builtin_metrics (Actuator/Micrometer)
PostgreSQL	—	—	exporter (postgres_exporter)
Redis	—	—	exporter (redis_exporter)
NGINX	—	—	exporter (nginx_exporter)

---

Workflow 2: Exporter Onboarding (Third-Party Systems)

Journey

Step-by-Step

Step	Action	Tool / Resource	Key Parameters
1	Browse catalog	Resource: prom://exporters/catalog	All exporters with ports, images, scopes
2	Get recommendation	prom_recommend_exporter(service_type="postgres")	Returns matching exporters + notes
3	Install exporter	prom_install_exporter(exporter_type="postgres_exporter", namespace="monitoring")	Creates Deployment + Service. MUTATES CLUSTER
4	Test endpoint	prom_test_endpoint(endpoint_url="http://postgres-exporter.monitoring:9187/metrics")	Validates Prometheus format
5	Apply ServiceMonitor	prom_apply_servicemonitor(namespace="monitoring", service_name="postgres-exporter")	Wires to Prometheus Operator
6	Verify end-to-end	prom_verify_exporter(backend_id="default", endpoint_url="...", job="postgres_exporter")	Polls endpoint + checks up{} series

Guided Prompt: Use prom-k8s-exporter-onboarding-guided for the full step-by-step flow.

---

Workflow 3: VM/Legacy Onboarding

For applications running on VMs or bare-metal servers. Uses file_sd_configs instead of ServiceMonitor.

Step	Action	Tool	Key Parameters
1	Recommend strategy	prom_recommend_instrumentation(environment="vm")	Returns VM-specific guidance
2	Deploy on VM	(Manual) — install exporter binary or instrumented app	—
3	Test endpoint	prom_test_endpoint(endpoint_url="http://10.0.1.5:8080/metrics")	Validates remote endpoint
4	Add to file_sd	prom_manage_file_sd(targets=["10.0.1.5:8080"], target_labels={"job": "my-app"})	Appends target + triggers /-/reload
5	Verify	prom_query_instant(query="up{job='my-app'}")	Confirms target is up

Guided Prompt: Use prom-vm-legacy-onboarding-guided for the full step-by-step flow.

---

Workflow 4: Synthetic Endpoint Monitoring (Probes)

For actively monitoring HTTP/TCP endpoint availability using Blackbox exporter + Probe CRDs.

Step	Action	Tool	Key Parameters
1	Install Blackbox	prom_install_exporter(exporter_type="blackbox_exporter", namespace="monitoring")	Default http_2xx module configured
2	Apply Probe	prom_apply_probe(targets=["https://talkops.ai"], probe_name="talkops-probe")	Auto-detects operator selector labels
3	Verify	prom_query_instant(query="probe_success")	Confirms synthetic check is running

---

Next Steps

• PromQL Querying — Safe query workflows with counter enforcement
• TSDB FinOps — Cardinality analysis and optimization
• Rule Management — Rule authoring, testing, and simulation
• Troubleshooting — Diagnosing failed scrape targets