Reference
Alert fields
| Field | Type | Meaning |
|---|---|---|
_id | Text | ALR- and the year and a number for your alerts (ALR-2026-000401); ALC- for common and ALP- for public alerts |
title | Text | Short description |
message | Text | What is wrong |
description | Text | More detail (optional) |
category | Text | A grouping, such as dataset, curve or system |
issue | Text | The type: delivery, completeness, quality, metadata, corrections, build, export or another |
severity | Text | The priority: critical, high, medium or low |
status | Text | open or closed |
agent | Text | Who it is with: client (your team), opendatadsl, or a data vendor's name |
subject | Object | What it is about: service (dataset, curve, data, process and so on), id, and ondate (the delivery or build date) |
key | Text | The id of the problem, used to count repeats instead of raising new alerts |
occurrences | Number | How many times it was raised while open |
lastOccurred | Date | When it was last raised |
tags | List | Your own labels |
assignee | Text | Who is working on it (optional) |
resolution | Text | What was wrong and what was done, set when closed |
created, updated, closed | Object | at (date) and by (who) |
relatedTo | Text | An earlier closed alert with the same key, if any |
root | Boolean | true on a root alert: the alert that explains what went wrong in a run (see Raising from a process) |
affects | List | On a root alert: the dataset ids the run was meant to load |
trace | Object | The run the alert was last raised in: id (the trace, shared by everything the run started), type (process or script), process (its name) and execution (its execution id) |
traces | List | Every trace id the alert has been raised under |
cause | Object | On an alert linked to a root cause (raised before it, or linked by hand): id and source of the root cause |
impactCount, openImpactCount | Number | On a root cause: the problems folded into it that you can see, and how many are still open |
impacts | List | On a root cause read by id: the problems folded into it (see below) |
linked | List | On a root cause read by id: the alerts linked to it, with _id, _source, title, status, severity and subject |
origin | Object | type: system (raised by the platform) or user |
commentCount | Number | Comments on the alert |
tenantCommentCount | Number | On shared alerts: your team's comments |
metrics | Object | Times in seconds: openSeconds, secondsClient, secondsOpenDataDSL, secondsVendor; reopens, handoffs, agentSince (when the current party got it) and firstResponseAt |
timeline | List | Everything that happened (see below). On shared alerts, your team's comments are only included when one alert is read |
environment | Text | Your alerts: the data environment |
_version | Number | Goes up with every change; send it back to make sure nobody changed the alert meanwhile |
Timeline entries
Every entry has id, type, at and by.
| Type | Extra fields | Meaning |
|---|---|---|
created | status, agent, severity | Raised |
comment | text; editedAt, deleted | A comment. On shared alerts your team's comments have scope: tenant and an id starting t- |
comment-edited | ref, text | A comment was edited (ref is the comment's id) |
comment-deleted | ref | A comment was deleted |
status | from, to | Closed or reopened |
agent | from, to | Handed over |
severity | from, to | Priority changed |
impact | impact, subject, issue, severity, via, title | On a root cause: a problem was folded in |
impact-resolved | impact, subject, issue, resolution | On a root cause: a problem folded in recovered |
linked | alert, alertSource, title on the root cause; root, rootSource, title on the alert | An alert was linked to the root cause: automatically (raised before the root cause) or by hand. A shared root cause does not name your alerts on its timeline |
unlinked | alert, alertSource, title on the root cause; root, rootSource on the alert | The link to a root cause was removed (or replaced by another) |
root-closed | root, rootSource, resolution | On a linked alert: its root cause was closed |
dataset-fixed | dataset, configSource, changes (target, field, from, to), outcome, text | OpenDataDSL corrected the dataset's configuration (calendar, expected tenors, times, priority or alerts) from this alert and checked the delivery again. When the delivery became a holiday or complete, the alert was closed |
Problems folded into a root cause
| Field | Meaning |
|---|---|
id | The problem's id |
subject | The item: service (dataset, curve or curve-export), id and ondate |
title, message, issue, severity, key | As the alert would have had them |
via | How it was linked: trace (same run), affects (named by the root cause), loader (loaded by its process), curve-input (built from an affected item) or export-curve (exports an affected curve) |
trace | The trace id of the run it came from, if any |
status | open or resolved (recovered), with resolvedAt and resolution |
occurrences, firstAt, lastAt | How many times it happened, first and last |
tid, environment | Whose item it is |
Who can do what
| Your alerts | Shared alerts | |
|---|---|---|
| See | Your company's alerts in the environment you work in | Every shared alert |
| Raise | Yes | No (OpenDataDSL only) |
| Comment | Yes | Yes, seen only by your company and OpenDataDSL |
| Close, reopen, hand over, change priority or details | Yes | No (OpenDataDSL only) |
| Edit or delete a comment | Its author, an administrator of your company, or OpenDataDSL | The same, for your company's comments |
| Delete an alert | Administrators of your company | No |
These rules are applied by the platform, whichever way you work: the screen, ODSL, Python or the REST API. Companies on the none plan do not have alerts.
What the extension installs
| Item | Id | What it does |
|---|---|---|
| Menu item | Operations > Alerts | The Alerts screen |
| Feature policy | extension.odsl.alerts | Who sees the menu item (everyone by default) |
| Security policy | AlertsPrivate | Read, raise, change and delete your alerts (category Alerts, source private, service alert) |
| Security policies | AlertsCommon, AlertsPublic | Read shared alerts and comment on them |
| Insight | odsl.alerts.audit | The Alert Audit Report, in the screen's Insights (category odsl.alerts) |
Without the security policies nobody in your company can see alerts, in the screen or through the API. To limit who works on alerts, change the members of these policies (for example to a group) in the portal's security settings.
The screen also reads your datasets (to know which you follow) and your dataset deliveries (for the delivery card), with your existing permissions on those services.