Prometheus Alert

Overview

The Prometheus Alert is a type of alert rule in Skylogs that connects to a Prometheus datasource or Alertmanager and retrieves the latest state of all alerts and their labels. You have strong label filtering to tune the rule to a specific alert and avoid generating noisy alerts.

A Prometheus alert rule matches alerts coming from a connected Prometheus / VictoriaMetrics data source. Alerts are ingested through the Prometheus & Alertmanager integration; this rule tells Skylogs which of those incoming alerts belong to it.

Before you start

1. Connect a Prometheus/VictoriaMetrics data source and confirm alerts are flowing in through Alertmanager.
2. Have at least one notification endpoint ready to attach.

Steps

1. Select the Alert Type → Prometheus.
2. Choose a Datasource — the Prometheus/VictoriaMetrics data source this rule should match against.
3. Choose a Query Type — Dynamic or Text Query (see below).
4. Enter a Unique Alert Name, assign Endpoints, and optionally add Tags.

Query types

Dynamic

Matches by the alert name as it already exists in the external Prometheus/VictoriaMetrics ruler — pick it from the connected data source rather than writing a query. Best when your alerting rules are already defined and maintained in Prometheus itself.

Field	Type	Description
dataSourceIds	array of string	Prometheus data source id(s) to match against
dataSourceAlertName	string	Alert name in the external Prometheus ruler, e.g. HighMemory
extraField	array of object	Optional label key/value filters to narrow the match

Text query

Matches using a raw PromQL expression, evaluated by Skylogs' own checker instead of relying on an existing Prometheus rule.

Field	Type	Description
queryText	string	PromQL expression string
queryObject	object	Structured query payload used by the checker

Example request

Dynamic:

{
  "type": "prometheus",
  "queryType": "dynamic",
  "name": "High memory usage",
  "description": "Fires when node memory exceeds threshold",
  "dataSourceIds": ["<prometheus-datasource-id>"],
  "dataSourceAlertName": "HighMemory",
  "endpointIds": ["<endpoint-id>"],
  "tags": ["production", "memory"]
}

Text query:

{
  "type": "prometheus",
  "queryType": "textQuery",
  "name": "High memory usage",
  "description": "Fires when node memory exceeds threshold",
  "queryText": "node_memory_MemAvailable_bytes / node_memory_MemTotal_bytes < 0.1",
  "endpointIds": ["<endpoint-id>"],
  "tags": ["production", "memory"]
}

Both are sent to POST /api/v1/alert-rule — see the full field reference in API reference → Alert rules.

Notes

• dataSourceIds can be left empty to match across all connected Prometheus sources.
• Use Dynamic when you want Prometheus's own ruler to remain the source of truth; use Text query when you want Skylogs to evaluate the condition itself.