Skip to content

Platform

Logs, console & metrics

Live logs in the browser, a shell in a running container, and CPU, memory and network metrics.

Every service page in the dashboard has three tabs for seeing what your container does: Logs, Console and Metrics.

Logs #

The Logs tab streams what your container writes to stdout and stderr, live. Write your app's logs there (not to files) to see them.

  • Follow all replicas or pick one, and pause the stream to read.
  • After a crash or restart, pick a replica and switch to its previous container to read what it printed before it stopped.
  • Logs come from the running container and the one before it. Older output is not kept, so a redeploy or a long sleep (scale-to-zero) starts the view fresh. Send logs to your own log service if you need history.
  • A stream reconnects on its own every 15 minutes; very chatty output is slowed down, never dropped.

Console #

The Console tab opens a shell inside the running container, in your browser — for checking files, running a migration or a one-off command.

  • Only the team owner can open a console. Every session is recorded in an audit log (who, when, which service), not its contents.
  • The image needs a shell: /bin/bash, /bin/sh or busybox. Minimal images (distroless, scratch) have none, so the console can't open there.
  • A session closes after 15 minutes without typing, and after 60 minutes at most.
  • Changes you make to files outside a volume are lost when the container restarts.

Metrics #

The Metrics tab charts, per replica:

  • CPU against your size's guaranteed CPU (a service may burst above it when the server has spare capacity) and memory against its limit — if memory hits the limit, the container is killed and restarted (an OOM kill), which the tab marks;
  • network in and out;
  • restarts, uptime and the last exit reason.

Pick the last hour, day or week; data is kept for 30 days. The same numbers are available from GET /v1/services/{id}/metrics and the get_service_metrics MCP tool — handy when you ask an AI assistant why a service keeps restarting.

Limits per plan #

PlanLive log streams at onceConsoles at once
Hobby 10 2
Pro 20 3
Enterprise 50 5

Something missing or wrong? Tell us — or open a ticket from the dashboard.