Alerts

Alert HTTP endpoints — global tab listings, per-entity listings, status mutation, halt-notification configuration, and the inbound AlertManager webhook.

Alerts surface platform-detected and externally-injected issues against catalog entities. The platform's own ingestion pipeline raises alerts for failed jobs, failed data-quality tests, backwards-incompatible schema changes, and distribution anomalies; the inbound AlertManager webhook (last group below) accepts Prometheus-style notifications and routes each one to a data entity by its entity_oddrn label. For the alert taxonomy, the per-tab UI (All / My Objects / Dependents), the lifecycle (OPEN / RESOLVED / RESOLVED_AUTOMATICALLY plus the auto-resolution triggers, the manual-reopen guard, and the housekeeping-driven cleanup), and the per-entity halt-notification feature, see the Alerting feature page; for outbound notification dispatch (Slack / email / generic webhook) and the inbound AlertManager wiring, see Notifications. Alert dispatch to Slack is platform-internal (driven by the WAL replication mechanism) and does not appear on this API surface — it's configured at notifications.receivers.slack.url and is distinct from the full Slack app used by Data Collaboration (datacollaboration.slack-oauth-token).

Global alert listings

The three tab queries on the navigation pane's Alerts section, plus a totals call used to render the badge counters. My Objects and Dependents require the signed-in user to be linked to an Owner; without the link, both endpoints respond with empty pages.

Method

Path

Operation ID

Purpose

GET

/api/alerts

getAllAlerts

Paginated list of every alert across the whole platform (back-end of the All tab).

GET

/api/alerts/my

getAssociatedUserAlerts

Paginated list of alerts on data entities where the signed-in user is a registered owner (back-end of the My Objects tab).

GET

/api/alerts/dependents

getDependentEntitiesAlerts

Paginated list of alerts on entities downstream (via lineage) of entities the signed-in user owns (back-end of the Dependents tab).

GET

/api/alerts/totals

getAlertTotals

Open-alert counts for each of the three tabs above; used to render the badge counters in the navigation pane.

Per-entity alert listings

Method

Path

Operation ID

Purpose

GET

/api/dataentities/{data_entity_id}/alerts

getDataEntityAlerts

Paginated list of alerts on a single entity (the entity-page Alerts tab).

GET

/api/dataentities/{data_entity_id}/alerts/counts?status={status}

getDataEntityAlertsCounts

Count of alerts on the given entity at the given status (OPEN / RESOLVED / RESOLVED_AUTOMATICALLY).

Alert status mutation

Method

Path

Operation ID

Purpose

PUT

/api/alerts/{alert_id}/status

changeAlertStatus

Resolve, auto-mark, or manually reopen an alert. Body: AlertStatusFormData carrying the target AlertStatus enum value. The activity feed records every transition as an ALERT_STATUS_UPDATED event.

Manual reopen has a guard. Setting the status back to OPEN is rejected with 400 Bad Request and the message Cannot reopen alert since the system already has an open alert of the same type if another alert of the same type is already open on the same data entity. Resolve or work the newer alert first, or leave the older one closed. The guard is enforced in AlertServiceImpl.updateStatus(...) (see AlertServiceImpl.java:124-131).

Per-entity halt-notification configuration

The halt configuration suppresses new alerts of a given type on a given entity until the configured timestamp passes. The toggle is per alert type and per entity; it is exposed on the entity's Notification Settings button in the UI and is also a thin REST surface. The platform records every change as an ALERT_HALT_CONFIG_UPDATED event on the Activity Feed (AlertHaltConfigServiceImpl.saveAlertHaltConfig is annotated @ActivityLog). For the operator-facing halt taxonomy, the five preset durations exposed on the UI, the suppress-only-creation semantics (auto-resolution still runs while halted), and the Distribution-anomaly known-limitation, see Alerting → Halt notifications per entity.

Method

Path

Operation ID

Purpose

GET

/api/dataentities/{data_entity_id}/alert_config

getAlertConfig

Read the current halt timestamps for the given entity. Returns a DataEntityAlertConfig with four fields (one per halt-able alert type — see below). Empty body means no halt is set.

PUT

/api/dataentities/{data_entity_id}/alert_config

updateAlertConfig

Write the halt timestamps for the given entity. Body: DataEntityAlertConfig. Pass an ISO-8601 date-time per field to set a halt; pass null for a field to clear it. Returns the updated config (201 Created).

The four halt-timestamp fields on DataEntityAlertConfig (all optional, all string ($date-time)):

Field

Suppresses

Notes

failed_job_halt_until

New Failed job alerts.

Halt is enforced on the ingestion-driven alert-creation path.

failed_dq_test_halt_until

New Failed data quality test alerts.

Halt is enforced on the ingestion-driven alert-creation path.

incompatible_schema_halt_until

New Backwards incompatible schema change alerts.

Halt is enforced on the ingestion-driven alert-creation path.

distribution_anomaly_halt_until

New Distribution anomaly alerts.

Currently unenforced — persisted by the API but ignored by the AlertManager-driven creation path. See Alerting → Distribution anomaly halt is currently unenforced.

Inbound AlertManager webhook

ODD also accepts Prometheus AlertManager notifications on a single inbound endpoint. Each inbound alerts[] entry is translated to a Distribution Anomaly alert on the data entity referenced by its entity_oddrn label. The endpoint is not covered by the OpenAPI spec (the AlertManager webhook payload is a Prometheus-defined contract) — the endpoint description, the accepted payload subset, the example AlertManager receiver configuration, and the operator-side authentication caveat (the endpoint is not behind the platform's ingestion auth filter) live with the operator-deployment context.

Method

Path

Purpose

POST

/ingestion/alert/alertmanager

AlertManager webhook receiver. Returns 204 No Content on success. The path is fixed in AlertManagerController.java:21.

For the payload subset accepted (alerts[].labels, alerts[].generatorURL, alerts[].startsAt), the example alertmanager.yml receiver configuration, the entity_oddrn label requirement, and the network-perimeter authentication options, see Configure ODD Platform → Prometheus AlertManager Integration.

See also