How do I surface service dependencies and on-call info in one place?

A good way to surface service dependencies and on-call info in one place is to create a service catalog / service portal that acts as the system of record for each service and pulls in data from your existing tools.

What to include

For each service, show:

• Owner team
• Primary on-call / secondary on-call
• Escalation policy
• Upstream dependencies and downstream dependents
• Runbooks / dashboards / alerts
• SLOs / health status
• Links to deploys, incidents, and docs

Common ways to implement it

1) Use an existing service catalog

Tools like:

• Backstage
• OpsLevel
• Compass
• Port
• Jellyfish (more delivery/engineering focus)

These can ingest:

• On-call schedules from PagerDuty, Opsgenie, VictorOps
• Dependencies from Kubernetes, cloud resources, CMDB, tracing/APM tools
• Ownership metadata from Git repos, YAML descriptors, service registries

This is usually the fastest path.

2) Build a lightweight internal portal

If you want something custom:

• Store service metadata in a repo or DB
• Sync on-call from PagerDuty/Opsgenie APIs
• Discover dependencies from:
  • service manifests
  • OpenTelemetry traces
  • Terraform / Kubernetes config
  • API gateway/service mesh data
• Render in a simple web UI with search and service pages

A common pattern is:

• Source of truth: repo-based service descriptor (catalog-info.yaml, service.yaml)
• Enrichment: API-synced on-call and dependency graphs
• Presentation: searchable portal with each service page

Recommended architecture

Best practice: keep ownership and static metadata in version control, and sync dynamic info from operational tools.

Example data sources:

• Service metadata: repo YAML/Backstage catalog
• On-call: PagerDuty API
• Dependencies: OpenTelemetry + service mesh + infra config
• Docs/runbooks: links from the service catalog

Good UX pattern

On each service page, show:

1. Service name and description
2. Current primary on-call
3. Escalation chain
4. Dependency graph
5. Recent incidents
6. Dashboards and alerts
7. Runbooks and owner team

If you want a simple starting point

Start with a table like:

Service	Owner	On-call	Upstream deps	Downstream deps	Runbook

Then evolve into a clickable service graph.

Practical recommendation

If you’re starting from scratch, Backstage + PagerDuty integration is one of the most common and effective solutions.

If you want, I can also give you:

• a minimal schema for service metadata
• a Backstage setup
• or a custom dashboard design for this use case.