LogForge Documentation
Browse documentation

Find and follow logs

Understand retained history, live logs and collection gaps.

LogForge · Runs in your infrastructure

Before you begin

  • Where: LogForge in your browser; no host shell needed.
  • Access: permission for the environment. Live viewing also needs Operate access and a connected agent.
  • Open: Overview > Hosts, expand the host, select a workload, then Logs.

Saved history stays available after you leave the page. Live output shows what is arriving now.

Logs controls with Search logs, Time range, Severity, Advanced, and a Live divider between saved and live rows.

Example workload. Saved rows appear above the Live divider.

Read saved history

  1. Choose a Time range covering the event.
  2. Use Search logs and Severity to narrow the rows.
  3. Open Advanced for a LogsQL filter, an expression that narrows stored queries.
  4. Clear filters when checking whether any data exists.

Check: the rows belong to the selected workload and time range. An empty result alone does not mean the agent is disconnected.

Retention: turn on Monitor logs and alerts to save new logs. Turning it off stops new retained logs and alert evaluation; previously saved evidence stays available.

Watch live output

  1. Choose the log source if the LogForge server offers more than one.
  2. Choose Start live logs if streaming has not already started.
  3. Watch new rows below the live boundary.
  4. Choose Stop when finished.

Check: new output arrives, then streaming stops. Buffered live rows can remain visible after Stop.

  • Monitored workload: the view can combine saved and live rows.
  • Live only · Not retained: new rows are not saved. Docker may supply a limited recent backlog.
  • View-only access: saved history may be available while live viewing is denied.

Leaving the page, signing out, or losing the connection ends the stream. Restore connectivity and start a fresh session; missed live output is not automatically replayed.

If logs are missing

Check application output, monitoring, filters, and time range. Docker collects standard output/error; a file inside the application is not automatically collected.

Follow No data, or continue with Metrics to check the second signal.