> ## Documentation Index
> Fetch the complete documentation index at: https://docs.bidsmith.pro/llms.txt
> Use this file to discover all available pages before exploring further.

# Monitoring

> Metrics, logs, traces, and alerts for ARIS nodes and registry services.

## What to monitor

<CardGroup cols={2}>
  <Card title="Availability" icon="heart-pulse">
    Node heartbeats, registry uptime, and request success rate.
  </Card>

  <Card title="Latency" icon="timer">
    P50/P95/P99 inference duration by model and node.
  </Card>

  <Card title="Cost" icon="wallet">
    Credits consumed per route, model, and tenant.
  </Card>

  <Card title="Capacity" icon="gauge">
    Queue depth, GPU utilization, memory pressure, and saturation.
  </Card>
</CardGroup>

## Recommended metrics

| Metric                        | Type      | Why it matters                        |
| ----------------------------- | --------- | ------------------------------------- |
| `aris_requests_total`         | Counter   | Total traffic per endpoint/model.     |
| `aris_request_latency_ms`     | Histogram | Detect slowdowns and regressions.     |
| `aris_inference_errors_total` | Counter   | Track node and model failures.        |
| `aris_queue_depth`            | Gauge     | Signal backpressure and scale events. |
| `aris_credits_deducted_total` | Counter   | Cost visibility and billing audits.   |

## Alert policy

<Tabs>
  <Tab title="Critical">
    Alert immediately when success rate drops below 97% for 5 minutes or registry is unreachable.
  </Tab>

  <Tab title="Warning">
    Alert when P95 latency rises 2x baseline for 15 minutes.
  </Tab>

  <Tab title="Capacity">
    Alert when queue depth exceeds planned concurrency for 10 minutes.
  </Tab>
</Tabs>

## Logs and trace correlation

Use a shared `request_id` across:

* client request logs
* registry routing logs
* node execution logs

This allows end-to-end failure triage in one query.