Use when checking service health, verifying infrastructure status, or after a deploy/restart. Checks containers, endpoints, Uptime Kuma, and Prometheus alerts.

Infrastructure Health Check

You are performing a fast, deterministic health check across all critical homelab services. This is not an exploratory debug session — it's a structured status sweep.

If an argument was provided, filter the checks to just that service. Otherwise, check everything.

Process

Step 1: Gateway Status Overview

Call gateway_call with sub-tool status via the Homelab Tools MCP. This returns connected service health (Qdrant, Neo4j, Postgres, n8n, Vault, Docker).

Step 2: Container Health

Call docker_call with sub-tool list_containers. Check these critical containers:

Container	Expected Status
homelab-mcp-gateway	running
pgvector-18	running
Neo4j	running
Qdrant	running
Vault	running
n8n	running
dify-api	running
dify-worker	running
dify-web	running
dify-plugin-daemon	running
dify_postgres17	running
openhands	running
Prometheus	running
Grafana	running
Alertmanager	running
uptime-kuma	running
Uptime-Kuma-API	running
RabbitMQ	running
Redis	running
MongoDB	running
gitlab-ce	running
traefik	running
docker-socket-proxy	running
ntfy	running

Flag any container that is:

Not running
Restarting
Started less than 2 minutes ago (possible recent crash-restart)

Step 3: Uptime Kuma Monitors

Call uptime_call with sub-tool list_monitors. Check for any monitor with status DOWN or PENDING.

Step 4: Prometheus Alerts

Call prometheus_call with sub-tool query with the query: ALERTS{alertstate="firing"}.

If any alerts are firing, list them.

Step 5: Key Endpoint Checks

Use remote-ssh to run quick curl checks from Tower:

curl -sf https://mcp.8-bit-byrum.com/health -o /dev/null && echo "gateway: OK" || echo "gateway: FAILED"
curl -sf http://dify-api:5001/v1/meta -o /dev/null && echo "dify-api: OK" || echo "dify-api: FAILED"
curl -sf http://Vault:8200/v1/sys/health -o /dev/null && echo "vault: OK" || echo "vault: FAILED"
curl -sf http://Prometheus:9090/-/healthy -o /dev/null && echo "prometheus: OK" || echo "prometheus: FAILED"
curl -sf http://Qdrant:6333/healthz -o /dev/null && echo "qdrant: OK" || echo "qdrant: FAILED"

Output Format

## Infrastructure Health Report

### Service Overview
| Service | Container | Status | Uptime | Notes |
|---------|-----------|--------|--------|-------|
| Gateway | homelab-mcp-gateway | UP | 3d 4h | |
| PostgreSQL | pgvector-18 | UP | 5d 1h | |
| Dify API | dify-api | DOWN | - | Stopped 10m ago |
| ... | ... | ... | ... | ... |

### Uptime Kuma Monitors
| Monitor | Status | Response Time |
|---------|--------|---------------|
| Gateway | UP | 142ms |
| ... | ... | ... |
(or "All monitors UP" if none are down)

### Prometheus Alerts
- No alerts firing
(or list of firing alerts)

### Endpoint Checks
| Endpoint | Status |
|----------|--------|
| gateway /health | OK |
| dify-api /meta | OK |
| vault /sys/health | OK |
| prometheus /-/healthy | OK |
| qdrant /healthz | OK |

### Summary
- Total services: 24
- Healthy: 24
- Degraded: 0
- Down: 0

If any service is down or degraded, add a Recommended Actions section with specific steps to investigate or fix.

infra-health