Guidal
Sign in

Metrics

CPU, memory and resource usage for services, databases and your whole account.

Metrics show how much of your allocated resources your workloads are actually using, over time. They answer the questions logs cannot: whether a service is near its memory limit, whether traffic is growing, whether a database is running out of room.

Service metrics

bash
guidal metrics service api

This reports CPU and memory usage for the service. Compare it against your plan's per-service allocation — a service consistently near its memory limit will eventually be restarted when it exceeds it.

Account-wide usage

bash
guidal metrics tenant

This shows consumption across everything you run, against your plan's limits. Use it to see how much headroom remains before a limit starts refusing operations.

Database metrics

bash
guidal db metrics demo-db

Connections and storage are the two to watch. Connections rising toward your pool size means requests will soon start waiting; storage approaching the limit means writes will soon start failing.

In the dashboard

  1. In the navigation pane, choose a service, database or Metrics.
  2. Choose the time range you want to inspect.
  3. Charts show usage against the relevant limit, so the ceiling is visible alongside the line.

What to watch

Memory approaching the limit. A container exceeding its memory allocation is terminated and restarted. Memory that climbs steadily and never falls is a leak; memory that spikes with traffic needs either a larger plan or a smaller working set per request.

CPU pinned at the limit. Sustained maximum CPU means requests are queueing. Either the work is genuinely heavy, or something is spinning.

Storage growth rate. Storage rarely matters until it is full, at which point writes fail. Watch the trend, not the current value — a database growing 1 GB a week on a 25 GB budget gives you months of warning if you look.

Restart count. Frequent restarts point at crashes, memory limits, or a failing health check.

Metrics and logs together

Metrics tell you that something changed; logs tell you why. A memory chart showing a step change at 14:00 is the question, and the logs from 14:00 are the answer. Start with metrics to find when, then read logs around that moment.