node: add metrics and observability

[ch4r10t33r]

Summary

The leanspec-node Docker image (currently published as ghcr.io/leanethereum/leanspec-node) runs the lean consensus specification as a node process, but has no instrumentation. Adding metrics and observability will make it possible to monitor node health, track protocol behaviour, and debug issues in devnet/testnet deployments.

Motivation

• Operators running leanspec-node have no visibility into internal state (slot progression, sync status, peer count, etc.)
• Without structured metrics there is no way to alert on anomalies or compare behaviour across multiple node instances
• Observability is a prerequisite for running a reliable multi-node devnet

Proposed Work

Metrics (Prometheus)

Expose a /metrics endpoint (default port 8008) with at minimum:

Metric	Type	Description
leanspec_slot_current	Gauge	Current slot number
leanspec_epoch_current	Gauge	Current epoch
leanspec_peers_connected	Gauge	Number of connected peers
leanspec_blocks_imported_total	Counter	Cumulative blocks imported
leanspec_attestations_processed_total	Counter	Cumulative attestations processed
leanspec_sync_distance	Gauge	Slots behind head (0 when synced)
leanspec_process_start_time_seconds	Gauge	Unix timestamp of node startup

Structured Logging

• Switch to structured (JSON) log output so log aggregators (Loki, Datadog, etc.) can parse fields without regex
• Include slot, epoch, and peer_id as log fields where relevant
• Log level controllable via --log-level CLI flag or LEANSPEC_LOG_LEVEL env var

Health Endpoint

Expose GET /health (same port as metrics, or a separate --rpc-port) returning:

{ "status": "ok", "slot": 12345, "synced": true }

Returns 200 when synced, 503 otherwise — suitable for Docker/Kubernetes readiness probes.

Docker / Deployment

• Expose port 8008 in the node stage of the Dockerfile
• Document the metrics and health endpoints in DOCKER_QUICKSTART.md
• Add a minimal docker-compose example with a Prometheus + Grafana sidecar

Acceptance Criteria

• /metrics endpoint returns valid Prometheus text exposition format
• /health endpoint returns correct synced field based on sync state
• Structured JSON log output when --log-format json is set
• Metrics port configurable via --metrics-port CLI flag
• Dockerfile node stage exposes the metrics port
• At least one example Grafana dashboard JSON included

[KatyaRyazantseva]

Should LeanSpec follow the Lean Ethereum metrics naming? Would be better.

[ch4r10t33r]

Fixed in #444