Open in ChatGPTOpen in Claude

APM monitoring with alerts

Create alerts straight from a service's APM metrics, so a rule inherits the query behind the chart you are looking at instead of being built from scratch. APM prefills the alert with a metric-threshold query scoped to the service, and the alerts - and the cases they open - surface back in APM: in the service's Monitoring tab and as columns in the catalog.

Use it to:

• Alert on what you see: turn any overview chart, on the catalog main screen or in a service or database drilldown, into a metric alert with the query prefilled.
• Keep alerts correlated with the service: prefilled APM labels tie each alert and its cases back to the service.
• Track status from the catalog: read a service's Alerts, Cases, and SLO status columns at a glance.
• Work the issue in one place: open the case behind a firing alert from the Monitoring tab.

What you need

• Coralogix Application Performance Monitoring (APM) installed and configured.
• Span Metrics enabled for the service.
• Permission to manage alerts. Creating alerts from an overview chart is available for services and databases.

Create an alert from a service's metrics

You can create an alert from any overview chart, both on the catalog main screen and inside a service or database drilldown. In a service Overview, the Requests, Errors, Latency, Latency by dependency, and Apdex cards each carry Create alert in their menu; on the catalog main screen, the same action sits on each overview widget. It is available for services and databases, and is hidden when a chart plots more than one query (for example, a latency chart showing several percentiles), since an alert takes a single query.

To create one from a service's Overview card:

1. In your Coralogix toolbar, select APM, then select a service to open its drilldown.
2. On the Overview tab, find a metric card: Requests, Errors, Latency, Latency by dependency, or Apdex.
3. Open the card's More actions menu () and select Create alert, then pick the metric to alert on.
4. The alert editor opens in create mode, prefilled as a Metric - Threshold alert. Its query is generated from the metric you picked, scoped to the current service, your selected environments, and the card's active group-by. You can keep the query or adjust it.
5. Define the remaining conditions and notification settings, then create the alert. The editor is part of Alerting: see that reference for the full alert definition.

Select the metric

Create alert opens a submenu of the service health metrics you can alert on:

• Avg latency
• Error %
• Latency (step), which opens a further choice of P50, P75, P90, P95, or P99
• Requests per minute
• Top time-consuming ops, which alerts on the top five operations' share of request time

Every metric produces a query scoped to the one service and the environments selected on the page. Top time-consuming ops groups by operation. The rest are ungrouped.

Alerts created this way carry APM labels - the product and the service name - so they appear in the Alerts screen with the service attached, and they surface back in the service's Monitoring tab.

You can start the same Create alert flow from several places, each prefilled for the entity and each offering the same metric submenu: an overview chart on the catalog main screen, a service's Overview metric card in the drilldown, its row in the catalog actions menu, the Service health hive, and a metric chart's series menu.

Note

You can also add service or database catalog labels to alerts you create outside APM. Those alerts are then classified as APM alerts and shown in the relevant APM surfaces. Don't modify the APM labels, or the alert loses its correlation with the service.

Track alerts, cases, and SLO status

In the catalog

The Catalog has three per-service columns:

• Alerts: the alerts scoped to the service.
• Cases: the cases open on the service. A firing alert opens a case.
• SLO status: the service's SLO health.

In the Monitoring tab

Open a service's Monitoring tab and switch to the Alerts & Cases panel to see every alert scoped to the service, enable or disable a rule inline, and open the case behind a firing alert.

Because a firing alert has already opened a case, there is no separate incidents view in APM - the alert row is the way in. Selecting a firing alert opens its case feed for the service; selecting any other alert opens the alert's drilldown. To work the incident in full, open the case from the alert row or go to Cases.

Next steps

See how a service interacts with its downstream systems in Dependencies.