o11y-full-setup - SKILL.md Agent Skill

name: o11y-full-setup description: Interactive guide for complete Observability setup — discovers data, presents summary, creates dashboards and SLOs based on user approval.

O11Y Full Setup Workflow

Guide the user through a complete Observability configuration for their Elastic deployment.

This is the umbrella skill. For focused tasks, see the related skills: o11y-slo-setup for SLO creation, o11y-service-dashboard for individual service dashboards.

Trigger

Use when the user asks to:

"Set up observability"
"Configure O11Y"
"Observability onboarding"
"Monitor my services"
"Full monitoring setup"

Also activates on keywords: "observability setup", "O11Y onboarding", "monitoring setup", "full observability", "end-to-end monitoring"

Do NOT use when:

"Create SLO" only (→ use o11y-slo-setup)
"Create dashboard" only (→ use o11y-service-dashboard)

Steps

0. Get Cluster Context

Call get_cluster_context to get cached cluster awareness — version, health, installed features, and observability capabilities. This determines what's already configured and what needs setup.

1. Discover Data

Call discover_o11y_data with no filters to get a complete picture of available data.

Present the results to the user as a summary:

How many APM services were found
How many hosts/containers are sending metrics
How many log sources exist
Data freshness status

2. Generate Summary with Recommendations

Call get_data_summary with the discovery results and format: "json".

Present the recommendations to the user:

"I found N dashboardable services and M SLO candidates"
List each recommended dashboard with its type and what it covers
List each recommended SLO with its indicator type and target

3. User Decision Point — Dashboards

Ask the user which recommended dashboards they want to create:

Present each as a numbered option
Allow "all" or specific selections
Allow the user to customize titles or skip entirely

4. Create Dashboards

For each approved dashboard, call kibana_api with POST /api/dashboards/dashboard and the recommended configuration as the request body. Report each created dashboard with its URL.

5. User Decision Point — SLOs

Ask the user which recommended SLOs they want to create:

Present each as a numbered option with service name, indicator type, and target
Allow "all" or specific selections
Allow the user to adjust targets before creation

6. Create SLOs

For each approved SLO, call kibana_api with POST /api/observability/slos and the configured parameters as the request body. Report each created SLO with its ID and URL.

7. Summary

Present a final summary:

Total dashboards created with URLs
Total SLOs created with IDs
APM instrumentation status
Log shipping status
Alert rules configured
Any warnings or issues encountered
Suggested next steps (add more services, configure additional alert rules, etc.)

8. Set Up Application Instrumentation (Optional)

If APM data is missing or the user wants to add new services, call setup_apm to guide application instrumentation:

Language-specific agent installation
Configuration for connecting to the Elastic APM server
Verification that data is flowing

9. Set Up Log Shipping (Optional)

If log data is missing or incomplete, call setup_log_shipping to configure log ingestion:

Identify log sources (application logs, system logs, container logs)
Configure Elastic Agent or Filebeat
Set up log parsing and field extraction

10. Create Alert Rules (Optional)

After dashboards and SLOs are created, offer to set up alerting using create_alert_rule:

Latency threshold alerts for critical services
Error rate spike alerts
SLO burn rate alerts
Infrastructure health alerts (CPU, memory, disk)

Tools Used

get_cluster_context — cached cluster awareness (version, health, capabilities)
discover_o11y_data — discover available O11Y data
get_data_summary — generate summary with recommendations
setup_apm — guide application instrumentation setup
setup_log_shipping — configure log ingestion
create_alert_rule — create alert rules for services and infrastructure
create_dashboard — high-level dashboard creation from discovered data
kibana_api — create dashboards (POST /api/dashboards/dashboard) and SLOs (POST /api/observability/slos)

API References

elastic://docs/api/kibana — Kibana REST API reference for dashboard and SLO endpoints

Prerequisites

ES_URL and ES_API_KEY (or basic auth) configured
At least some O11Y data flowing into Elasticsearch

Related Skills

o11y-slo-setup — focused SLO creation workflow
o11y-service-dashboard — focused service dashboard creation