Metrics and logs
See CPU, memory, and network usage, stream deployment logs, and open a shell in a running service.
Each service panel has tools for seeing what a service is doing.
Metrics
Open the service panel → Metrics tab.
| Chart | What it shows |
|---|---|
| CPU | vCPU used over time |
| Memory | Memory used over time |
| Network | Data in and out |
Use the range selector to switch between the last hour, day, week, or month. Readings are taken every minute and kept for 30 days.
Metrics are the fastest way to answer two questions:
- Is the service sized right? If memory sits close to the limit, raise it before the service starts getting killed and restarted. If CPU never rises above a fraction of what you allocated, lower it and pay less.
- Is a spike load or a leak? Load goes up and comes back down. A leak only goes up, and resets when the service restarts.
Deployment logs
Open the service panel → Deployments tab → View Logs on a deployment.
Logs stream live. You can filter by text to narrow them down, and open logs for an older deployment from the history list.
Deployment logs are not available for Postgres services. Use the Metrics tab and the parameters instead.
Shell access
Open the service panel → Shell tab to get a terminal inside a running Docker service. Use it to check a file exists, look at an environment variable, or run a one-off command.
A shell session runs inside one instance of the running service. Anything you write to disk is lost on the next deploy or restart, and is not shared with other instances. Never use it to make a permanent change — change the image, the config mounts, or the variables instead.
Network traffic
The Network tab shows connections in and out of the service — which addresses it talked to, on what ports, and how much data moved.
This is useful for finding out what a service is actually reaching. A service talking to the public internet when you expected it to stay internal will show up here.
