Skip to content

Logs and diagnostics #

Logs are the first line of diagnosis when a device misbehaves in the field. If devices are already deployed, start with one representative connected unit and prove a test event now. Add the same configuration to the product image before the next shipment or reimage.

Why device-tied logs matter #

Generic log platforms collect streams of data but often lack device identity. When support gets a ticket about "the gateway at site 7," they need logs tied to that specific device - not a search across unrelated streams.

Setup approach #

  1. Identify the log sources that matter for your application (application logs, systemd journals, custom output).
  2. Use the current Logs onboarding panel to select the supported source and obtain its organisation-specific instructions.
  3. Configure one representative test device or producer.
  4. Send and find a distinctive test event.
  5. Include the proven configuration in your golden image or existing deployment process.
  6. Teach responders the search path before the first incident.

Device workspace evidence #

After devices ship, operators usually need more than the log stream:

SurfaceUse it for
LoggingOrganisation-retained application and system lines
Device DiagnosticsHost and system facts on the device record
Device timelineRecent lifecycle and operational events
Agent LogDataplicity agent output for that unit

If Diagnostics is available to your organisation, use it to inspect device hardware and system information before opening a shell.

When to use this #

  • Application errors need investigation without SSH
  • Support wants context before opening a remote shell
  • Engineering needs visibility into field behaviour