Telemetry: Prometheus metrics + health endpoints (4‑PR series)

[Jim8y]

Summary

This issue tracks the introduction of first‑class node telemetry for Neo N3 via a new Prometheus‑exporting plugin, health endpoints, and accompanying monitoring assets.

Motivation

Node operators currently rely on ad‑hoc log inspection and RPC polling for visibility. A built‑in telemetry surface enables:

• consistent operational monitoring across deployments
• alerting on sync/peer/mempool/resource health
• easy integration with existing Prometheus/Grafana stacks

Scope / PR stack

The work is delivered as four small, stacked PRs targeting the telemetry staging branch:

1. Core hooks – [Telemetry 1/4] Add telemetry hooks to Neo core #4372 (telemetry-1)

   • Adds non‑intrusive P2P events in RemoteNode (MessageSent, ConnectionChanged) for plugins to observe network activity.
   • Exposes Plugin.IsStopped as protected internal so plugins can self‑disable safely.

2. Telemetry plugin & metrics – [Telemetry 2/4] Add Telemetry plugin with metrics collection #4373 (telemetry-2)

   • Introduces the Neo.Plugins.Telemetry plugin with collectors for blockchain, network, mempool, system, and plugin state.
   • Exposes a Prometheus endpoint (default http://localhost:9100/metrics).
   • Uses consistent labels: node_id, network.

3. Health endpoints – [Telemetry 3/4] Add health check endpoints for telemetry #4374 (telemetry-3)

   • Adds lightweight HTTP /health, /ready, and /live endpoints.
   • Readiness reflects both sync status and peer connectivity.

4. Docs & monitoring assets – [Telemetry 4/4] Add telemetry documentation and monitoring configs #4375 (telemetry-4)

   • Comprehensive README with install/config/metric reference.
   • Prometheus scrape snippet and alert rules.
   • Grafana dashboard JSON for node operators.

PRs should be merged in order 1 → 4. After that, the telemetry branch can be fast‑forwarded into the target release branch.

Usage

Enable the plugin via Plugins/Telemetry/config.json, then configure Prometheus to scrape:

scrape_configs:
  - job_name: neo-node
    static_configs:
      - targets: ['<host>:9100']

Health checks:

• GET /health
• GET /ready
• GET /live

Validation

• The plugin builds cleanly at each PR boundary.
• Collectors are guarded so telemetry cannot destabilize the node unless explicitly configured with ExceptionPolicy = StopNode.

Follow‑ups

• Add unit/integration tests for collectors once a Telemetry test harness is agreed.
• Consider more granular per‑peer metrics and optional OpenTelemetry traces.

[vncoelho]

As I commented before in the first PR, I did not understand why you are keep pushing this plugin here instead of in Neo-Node folder, as we migrated all plugins to that repo now, @Jim8y.

In any case, I think it should be there. Here in this repo you should focus only on any method that eventually may need some new exposure.