render-monitor

Run a five-step health check: list services for status, check the latest deploy, pull recent error logs, and query CPU, memory, and p95 HTTP latency metrics. Healthy indicators are a live deploy, error rate under 0.1%, and p95 latency under 500ms.

Use the list_logs MCP tool with the service ID, filtering by level such as error, by status codes like 500 or 503, by search text, or by time range. If MCP is unavailable, use the CLI command render logs with the service ID and level flags.

Available metrics include cpu_usage, memory_usage, http_latency at chosen quantiles, http_request_count, bandwidth_usage, instance_count, and active_connections for databases. Each metric has documented healthy, warning, and critical thresholds in the metrics reference guide.

Partially. The Render CLI can handle service status, deployments, and logs as a fallback, but metrics queries and direct PostgreSQL queries require the Render MCP server configured with an API key.

Use the query_render_postgres MCP tool to run SQL against pg_stat_statements, ordering by mean execution time to surface the slowest queries. You can also check active connection counts against your plan's limits.

High CPU often comes from inefficient code or missing caching, while high memory can indicate leaks or large in-memory data. Compare usage against the plan limit to assess throttling or OOM risk, then optimize queries, add caching, or scale the plan.