Working with alerts
You can work on alerts in the portal, in code, or let automations act on them for you.
Ways in​
| How | What you can do | Documentation |
|---|---|---|
| The portal | Operations > Alerts: your alerts and the shared alerts you follow, with filters and figures; one alert with its dataset delivery, root cause and timeline; raise, comment, hand over, change the priority, close and reopen | The Alerts extension, The Alerts screen, Working with alerts |
| ODSL | Find, read, raise and change alerts with the alert service | Alerts in ODSL |
| Python | The same with the Python SDK | Alerts in Python |
| REST API | alert/v1/private, common and public | Alerts REST API |
| Fusion AI | Ask an assistant to raise an alert | Where alerts come from |
Notifications​
To hear about alerts as they happen, add an automation on the alert service. Its actions are:
| Action | When |
|---|---|
raised | An alert is raised, or a problem is folded into a root cause |
closed | An alert is closed, or a folded problem recovers |
reopened | An alert is reopened |
agent | An alert is handed over |
severity | An alert's priority changes |
comment | A comment is added, edited or deleted |
The condition's id is what the alert is about (a dataset or curve id) or * for every alert, so "email the gas desk when anything about ICE.NDEX.TTF is raised" is one automation. See Notifications and automations for examples, and Automation basics for how automations work.
Insights​
| Insight | What it answers |
|---|---|
| Alert Audit Report | Who raised, commented on, handed over and closed your alerts, and when; the whole history of one alert; how quickly alerts were answered and closed |
| Tracer | Everything one run started, by its trace id |
Where next​
- The Alerts extension
- Alerts reference: every field of an alert and its timeline entries