Logs
The device Observe > Logs page searches the logs from one device: kernel messages, Admiral platform logs, and your workload's stdout and stderr. Fleet > Observe > Logs does the same across every device in a fleet.

Searching
- Pick a time range at the top right (default 6 hours). A histogram shows when matching lines occurred.
- Level: filter by ERROR, WARN, INFO or DEBUG.
- Source: the sources this device has reported, such as the kernel, the platform, and your workload.
- Search: free text, or click the code button to write a full LogsQL query (
Ctrl/Cmd + Enterruns it). - Choose how many rows to return (default 500). Click a row to see all of its fields.
Live
Live streams new matching lines as they arrive; Stop ends the stream. Long-running streams reconnect automatically every 15 minutes.
If a device loses its connection, it keeps its logs on disk and sends them when it reconnects, so the history has no gaps.
Next Steps
- Device State & Diagnostics for a one-glance diagnosis and live probes.
- The
/sys/admiralinterface for reading live device status from inside your workload.