Observability

"Born with Eyes." — Every A.R.C. platform ships with enterprise-grade observability from day one.

Overview

A.R.C. provides a complete, pre-configured observability stack. When you enable the observability feature flag, six services activate to give you full visibility into your platform:

Your Services
     │
     ▼
Black Widow (OpenTelemetry Collector)
     │
     ├──→ Dr. House (Prometheus) ──→ Metrics
     ├──→ Watson (Loki) ←── Hermes (Promtail) ──→ Logs
     └──→ Columbo (Jaeger/Tempo) ──→ Traces
                    │
                    ▼
            Friday (Grafana)
          "The Single Pane of Glass"

The Observability Stack

Service	Codename	Role
OpenTelemetry Collector	Black Widow	The single entry point for all telemetry signals
Prometheus	Dr. House	Metrics collection and alerting — "It's never DNS"
Loki	Watson	Log aggregation — writes down every detail for later
Jaeger/Tempo	Columbo	Distributed tracing — "Just one more thing"
Grafana	Friday	Unified dashboards — the visual interface for everything
Promtail	Hermes	Log shipping — delivers logs from containers to Watson

Enabling Observability

In your arc.yaml:

features:
  observability: true

Then run:

arc workspace run

Resource impact: +4GB RAM, +2 CPU cores (6 additional services).

Key Dashboards

Once enabled, access Friday (Grafana) at http://localhost:3000:

• Service Health: Overview of all service status and uptime
• Request Tracing: End-to-end request paths across services
• Log Explorer: Search and filter logs from all services
• Metrics Explorer: Custom queries against all collected metrics

The Three Pillars

A.R.C. implements all three pillars of observability:

1. Logs (Watson) — What happened, in detail
2. Metrics (Dr. House) — How the system is performing, numerically
3. Traces (Columbo) — How requests flow through services
4. Dashboards (Friday) — Unified visualization of all signals

---

Explore each component in detail using the sidebar.