Skip to main content

Lanternfish docs

Build from first trace to first briefing

Browse documentation
Core concepts

Concept

Evidence

The runs behind an insight, so every number can be checked against the runs it came from.

How runs are ranked

  • Primary: the runs linked to the insight, the ones it was measured on.
  • Supporting: other successful runs from the same window.
  • Reference: other failed runs from the same window, kept for comparison.

How they are summarized

Runs are grouped by status and by the three most frequent agents. The overview names the agent that appears most often and the deployment recorded in that window, if any. The first two primary runs and the first supporting run are shown first; the rest stay one click away.

Next: Narrative