Metrics Specification

Prometheus-compatible metrics exposed by the Stella Ops Release Orchestrator.

Audience: SREs and platform operators building dashboards, SLOs, and alerts for the Release Orchestrator. Status: Planned — this is the reference metrics catalog; treat the names here as the canonical contract. Related: Alerting · Logging · Tracing · Operations Overview

Overview

The Release Orchestrator exposes Prometheus-compatible metrics for monitoring deployment health, performance, and operational status. Metrics are grouped by domain (releases, promotions, deployments, agents, workflows, targets, integrations, gates) plus cross-cutting API and evidence series. All metric names are prefixed stella_. The same metrics back the alert rules in Alerting and the SLIs defined below.

Core Metrics

Release Metrics

Promotion Metrics

Deployment Metrics

Agent Metrics

Workflow Metrics

Target Metrics

Integration Metrics

Gate Metrics

API Metrics

Evidence Metrics

Prometheus Configuration

# prometheus.yml
global:
  scrape_interval: 15s
  evaluation_interval: 15s

scrape_configs:
  - job_name: 'stella-orchestrator'
    static_configs:
      - targets: ['stella-orchestrator:9090']
    metrics_path: /metrics
    scheme: https
    tls_config:
      ca_file: /etc/prometheus/ca.crt

  - job_name: 'stella-agents'
    kubernetes_sd_configs:
      - role: pod
        selectors:
          - role: pod
            label: "app.kubernetes.io/name=stella-agent"
    relabel_configs:
      - source_labels: [__meta_kubernetes_pod_label_agent_id]
        target_label: agent_id

Histogram Buckets

Duration Buckets (seconds)

# Short operations (API calls, gate evaluations)
short_duration_buckets: [0.01, 0.025, 0.05, 0.1, 0.25, 0.5, 1, 2.5, 5, 10]

# Medium operations (workflow steps)
medium_duration_buckets: [0.1, 0.5, 1, 2.5, 5, 10, 30, 60, 120, 300]

# Long operations (deployments)
long_duration_buckets: [1, 5, 10, 30, 60, 120, 300, 600, 1200, 3600]

Size Buckets (bytes)

# Request/response sizes
size_buckets: [100, 1000, 10000, 100000, 1000000, 10000000]

# Evidence packet sizes
evidence_buckets: [1000, 10000, 100000, 500000, 1000000, 5000000]

SLI Definitions

Availability SLI

# API availability (99.9% target)
sum(rate(stella_http_requests_total{status!~"5.."}[5m]))
/
sum(rate(stella_http_requests_total[5m]))

Latency SLI

# API latency P99 < 500ms
histogram_quantile(0.99,
  sum(rate(stella_http_request_duration_seconds_bucket[5m])) by (le)
)

Deployment Success SLI

# Deployment success rate (99% target)
sum(rate(stella_deployments_total{status="succeeded"}[24h]))
/
sum(rate(stella_deployments_total[24h]))

Alert Rules

groups:
  - name: stella-orchestrator
    rules:
      - alert: HighDeploymentFailureRate
        expr: |
          sum(rate(stella_deployments_total{status="failed"}[1h]))
          /
          sum(rate(stella_deployments_total[1h])) > 0.1
        for: 5m
        labels:
          severity: critical
        annotations:
          summary: High deployment failure rate
          description: More than 10% of deployments failing in the last hour

      - alert: AgentOffline
        expr: stella_agent_heartbeat_age_seconds > 120
        for: 2m
        labels:
          severity: warning
        annotations:
          summary: Agent {{ $labels.agent }} offline
          description: Agent has not sent heartbeat for > 2 minutes

      - alert: PendingApprovalsStale
        expr: |
          stella_approval_pending_count > 0
          and
          time() - stella_promotion_request_timestamp > 3600
        for: 5m
        labels:
          severity: warning
        annotations:
          summary: Stale pending approvals
          description: Approvals pending for more than 1 hour

      - alert: IntegrationUnhealthy
        expr: stella_integration_health == 0
        for: 5m
        labels:
          severity: warning
        annotations:
          summary: Integration {{ $labels.integration }} unhealthy
          description: Integration health check failing

      - alert: HighAPILatency
        expr: |
          histogram_quantile(0.99,
            sum(rate(stella_http_request_duration_seconds_bucket[5m])) by (le, path)
          ) > 1
        for: 5m
        labels:
          severity: warning
        annotations:
          summary: High API latency on {{ $labels.path }}
          description: P99 latency exceeds 1 second

Grafana Dashboards

Main Dashboard Panels

1. Deployment Pipeline Overview

   • Promotions per environment (time series)
   • Success/failure rates (gauge)
   • Active deployments (stat)

2. Agent Health

   • Connected agents (stat)
   • Agent status distribution (pie chart)
   • Heartbeat age (table)

3. Gate Performance

   • Gate evaluation counts (bar chart)
   • Block rate by gate type (time series)
   • Evaluation latency (heatmap)

4. API Performance

   • Request rate (time series)
   • Error rate (time series)
   • Latency distribution (heatmap)

References

• Operations Overview
• Logging
• Tracing
• Alerting