docs: document new logs and prometheus metrics

docs/admin/integrations/prometheus.md

@@ -102,6 +102,27 @@ You must first enable `coderd_agentstats_*` with the flag
 `CODER_PROMETHEUS_COLLECT_AGENT_STATS` before they can be retrieved from the
 deployment. They will always be available from the agent.
 
+### Recent Metrics Enhancements
+
+Several metrics have been enhanced with additional labels to provide more detailed monitoring:
+
+- **`coderd_api_concurrent_requests`**: Now includes `method` and `path` labels to track concurrent requests per API endpoint and HTTP method
+- **`coderd_api_concurrent_websockets`**: Now includes a `path` label to track concurrent websocket connections per API endpoint
+- **`coderd_api_requests_processed_total`**: Includes special handling for route patterns:
+  - Routes not matching a specific pattern are labeled as "UNKNOWN"
+  - Static file requests are labeled as "STATIC"
+
+These enhancements allow for more granular monitoring and better insight into API performance by endpoint.
+
+### Deprecated Metrics
+
+The following metrics have been deprecated and replaced with metrics that follow Prometheus naming conventions:
+
+- **`coderd_api_workspace_latest_build_total`**: Replaced by `coderd_api_workspace_latest_build` (gauge metrics should avoid the `_total` suffix)
+- **`coderd_oauth2_external_requests_rate_limit_total`**: Replaced by `coderd_oauth2_external_requests_rate_limit`
+
+Please migrate to the new metric names in your dashboards and alerts.
+
 <!-- Code generated by 'make docs/admin/integrations/prometheus.md'. DO NOT EDIT -->
 
 | Name                                                          | Type      | Description                                                                                                                      | Labels                                                                               |

docs/admin/monitoring/logs.md

@@ -22,6 +22,16 @@ machine/VM.
 Events such as server errors, audit logs, user activities, and SSO & OpenID
 Connect logs are all captured in the `coderd` logs.
 
+### Request Logs
+
+Request logs include detailed information to help troubleshooting and auditing:
+
+- **Authentication Context**: Logs include requestor ID, name, and email for user subjects
+- **RBAC Subject Types**: Clearly identifies system services (provisioners, autostart, etc.)
+- **Route Parameters**: Includes workspace and agent name parameters for better traceability
+- **URL Parameters**: Query parameters are included with the prefix `params_`
+- **Database Context**: Information about database operations related to the request
+
 ## `provisionerd` Logs
 
 Logs for [external provisioners](../provisioners/index.md) are structured

docs/admin/monitoring/metrics.md

@@ -20,3 +20,42 @@ links point to relevant sections there.
   (Kubernetes users only)
 - [Configure Prometheus to scrape Coder metrics](../integrations/prometheus.md#prometheus-configuration)
 - [See the list of available metrics](../integrations/prometheus.md#available-metrics)
+
+## Monitoring Best Practices
+
+### Key Metrics to Monitor
+
+Here are some of the most important metrics to track for operational health:
+
+1. **API Performance**
+   - `coderd_api_request_latencies_seconds`: Track latency by endpoint to identify slow APIs
+   - `coderd_api_concurrent_requests`: Monitor by path and method to identify bottlenecks
+   - `coderd_api_requests_processed_total`: Track error rates by status code
+
+2. **Workspace Status**
+   - `coderd_workspace_latest_build_status`: Monitor workspace build success rates
+   - `coderd_agents_connections`: Track agent connectivity
+
+3. **User Activity**
+   - `coderd_api_active_users_duration_hour`: Monitor active user count
+   - `coderd_insights_templates_active_users`: Track template usage
+
+4. **System Health**
+   - `go_memstats_alloc_bytes`: Monitor memory usage
+   - `process_cpu_seconds_total`: Track CPU utilization
+
+### Example Prometheus Queries
+
+These queries can help monitor common scenarios:
+
+```
+# API Error Rate (5xx errors as percentage of total)
+sum(rate(coderd_api_requests_processed_total{code=~"5.."}[5m])) / 
+sum(rate(coderd_api_requests_processed_total[5m])) * 100
+
+# Slow API Endpoints (95th percentile latency > 1s)
+histogram_quantile(0.95, sum(rate(coderd_api_request_latencies_seconds_bucket[5m])) by (path, le)) > 1
+
+# Failed Workspace Builds (last 24h)
+sum(increase(coderd_workspace_builds_total{status="failed"}[24h])) by (template_name)
+```