Run your app
Logs and metrics
The Logs tab#
Logs shows console output from the service's running containers, across the servers it runs on. Container Apps and services with several copies include container names; hover over one to see its server. Build output belongs in Builds, and rollout progress in Deployments.
| Control | What it does |
|---|---|
| Following / Follow | Following streams new lines as they arrive. Click it to stop following; use Follow to resume. |
| Start from | Choose Last 100 lines, Last 500 lines, Last 15 minutes, Last hour or Last day. |
| Filter lines | Search the loaded messages and container names, ignoring letter case. |
| Clear | Clears the view and starts again from the selected starting point. Does not delete the app's logs. |
| Jump to latest | Returns to the newest output after you scroll up. |
A recent view, not a log archive
The screen keeps at most 2,000 lines, dropping the oldest as new ones arrive. Each container keeps roughly 30 MB of rotated output, rather than a fixed number of days, so choosing Last day does not guarantee a full day is available. Rotated logs and output from replaced containers may no longer be available. A service with no running containers has no runtime logs to show.
The view reconnects after a dropped connection or a stream time limit, resuming from the last received line.
The Metrics tab#
Metrics shows measured container CPU and memory, added across copies. CPU is in cores: 1 core means one CPU fully busy. Memory excludes reclaimable file cache.
| Time range | Chart points |
|---|---|
| 24 hours | 5-minute averages and peaks. |
| 7 days | Hourly averages and peaks. |
| 30 days | Hourly averages and peaks. |
- Size shows the app's CPU and memory allowance, whether limits are applied, peak memory in the chosen window, out-of-memory events in the last 24 hours and temporary disk use where reported.
- Memory and CPU charts show average and peak usage.
- By component splits a Container App's usage by its parts. Switch between Memory and CPU; the table shows Now, Average, Usually under (p95) and Peak. The chart groups components beyond the first seven into Other.
Figures depend on the server reporting samples. Missing measurements are not zero usage; the page says when there are no samples. The dashboard offers history up to 30 days. Use Sizes and limits to decide whether to change the app's size.