Monitoring Runs

The Runs page is your real-time window into what your agents are doing. Every run tells the complete story of an agent's task execution — from trigger to outcome.

The Runs Page

Navigate to Runs in the sidebar to see all runs across your workspace. Each row shows:

• Agent name — which agent executed the run
• Trigger — what initiated the run
• Status — running, completed, errored, or escalated
• Actions — how many actions were requested during the run
• Started at — when the run began
• Duration — how long the run took

Runs are listed in reverse chronological order. Active runs appear at the top with a live indicator.

Run States

Inspecting a Run

Click on any run to open its detail view. The detail view shows the complete run story in a single page:

• Run metadata (agent, trigger, duration, outcome)
• Full event timeline
• Action requests and their results
• Policy evaluations
• Execution details

Event Timeline

The timeline shows every event logged during the run, in chronological order. Each event includes:

• Timestamp
• Event type (e.g. context.retrieved, intent.classified)
• Event data (structured JSON with the details)

The timeline is the most important debugging and auditing tool in Koladr. It lets you reconstruct exactly what the agent did, what data it used, and what decisions it made.

Policy Evaluation Visibility

For every action in a run, the detail view shows which policies were evaluated and what the outcome was. This includes:

• Which policies matched the action
• What effect was applied
• If an approval was required, who approved or rejected it

This transparency makes it easy to understand why an action was allowed, blocked, or sent for approval.

Execution Results

When an action is executed through a connector, the run timeline includes the execution result. This shows:

• Whether the execution succeeded or failed
• The response from the external service
• Execution time
• Any error details if the execution failed

Filtering and Search

Use the filters at the top of the Runs page to narrow your view:

• Agent — show runs from a specific agent
• Status — filter by run state (running, completed, errored)
• Date range — focus on a specific time period

Audit and Replay

Every run is a complete, immutable record. This makes runs invaluable for:

• Compliance audits — demonstrate exactly what your agents did, when, and why
• Debugging — trace through the event timeline to find where something went wrong
• Policy tuning — review past runs to identify actions that should have been blocked or approved
• Performance analysis — understand how long different types of runs take and where bottlenecks occur