Runs

What it is for

One list of what every agent actually did, newest first. When the dashboard shows the success rate dropping, this is where you find out which runs failed and why.

How to reach it

Runs in the navigation, between Approvals and Memory.

A link in a Talk delivery or a failure alert opens the same page, already narrowed to the schedule that produced it. The page says so, and offers to show you everything.

What you see

Column	What it tells you
When	The time the run was recorded
Agent	Which agent ran, linked to its detail page
Status	How it ended
Trigger	Whether a schedule started it or a flow did
Duration	How long it took
Summary	What the run reported

Filter by agent, by status, or both. The list pages fifty at a time and tells you the total, so "1 to 50 of 340" means what it says.

Why the Trigger column exists

Hermiq starts an agent two ways, and they record against different objects. A scheduled run hangs on its schedule. A flow-triggered run hangs on whatever object the flow was about, which usually lives in a different register entirely.

Before this page, no list could show you the second kind. The dashboard counted them and every listing denied they existed. Naming the channel on each row keeps that honest.

What is not listed

Dry runs and replays. They are previews, and a preview that appears in the history is indistinguishable from something that happened.

A run whose agent was deleted also disappears, because visibility is resolved through the agent. That is a known gap in the OpenRegister read path, stated here rather than hidden.

API

• GET /apps/hermiq/api/runs lists runs, with agentId, status, limit, offset
• GET /apps/hermiq/api/schedules/{scheduleId}/runs lists one schedule's runs
• GET /apps/hermiq/api/schedules/{scheduleId}/runs/{runId}/trace returns the step timeline

Where to go next

Found a failing agent? Open it from the Agent column and read its run history and memory in place.