Skip to content
Agirunner

Diagnostics

Diagnostics is where operators move from workflow state to execution evidence.

This is a crucial difference between a demo and an operating surface. Agirunner is designed so that operators can answer “what happened?” with actual evidence rather than with guesswork.

For a quick stack identity check, use the Versions popover above Logout in the sidebar. It gives you a compact read of the running platform, dashboard, container-manager, and runtime image versions without leaving the current page.

Live Logs

Section titled “Live Logs”

The logs surface supports both:

  • raw stream inspection
  • a summarized activity view over the filtered result set

That lets an operator stay close to the source of truth without giving up higher-level signal.

The current implementation also supports:

  • workflow-scoped inspection
  • filter controls around operation, role, and actor
  • summary cards over the current filtered result set
  • workflow budget context when the operator is inspecting one workflow in detail

Live Containers

Section titled “Live Containers”

The live container inventory shows:

  • orchestrator-side containers
  • specialist-side containers
  • current and recently active execution posture

This helps answer whether the runtime fleet is actually doing work, not just whether the control plane believes it is.

Why Diagnostics Matter

Section titled “Why Diagnostics Matter”

An agent system that cannot explain what happened is hard to trust. Diagnostics are first-class in Agirunner because operators need evidence, not vibes.

How It Connects To The Rest Of The System

Section titled “How It Connects To The Rest Of The System”

Diagnostics is what makes the rest of the dashboard trustworthy:

The workflow surfaces tell you what moved. Diagnostics helps answer why it moved, why it stalled, or why the runtime posture does not match the control-plane story.