Getting started
Step 4 of 5
Read your first briefing
The briefing tells you what changed in your agents over the last 24 hours, what each change is compared with, and where to look next.
What is on the page
- The lead sentence says how many changes need your attention, followed by the number of runs in the last 24 hours and when the briefing was last updated.
- Three numbers: active users, successful runs and failure rate. Each arrow compares with the previous 24 hours.
- The changes, most serious first. Each one states its numbers and basis, for example: Flights api average latency is up 27%: 432 ms vs 340 ms in the previous 7 days. Based on 212 calls in the last 24 hours vs 1,840 calls in the previous 7 days.
- Recent events: deployments and other events from the last hours, so you can see what happened just before a change.
When nothing has changed, the briefing says so. When there are no runs yet, it shows how to send the first one instead of empty charts.
How to read the priority
| Critical | The change passed the critical threshold, for example latency up 35% or more. Look at it first. |
|---|---|
| Warning | A clear change in the bad direction, for example latency up 20% or more. |
| Improvement | A smaller change that passed the first threshold; worth a look when you have time. |
The thresholds for each metric are listed under Insight. A change that improves things is not raised as a problem.
How it is built
Every item is computed from your runs. The short summary above the items is written from them by a language model and uses one AI briefing from your plan; when none are left, you see a plain summary built from the same items. The briefing is rebuilt at most every 30 minutes.
Next
Open any change to see its investigation: the numbers behind it, steps worth checking and the runs that show it.