Monitoring with Alerts

Manage alerts within your APM interface to monitor critical metrics and receive timely notifications about potential issues.

Overview

Metric-based widgets in your Application Performance Monitoring (APM) dashboard are crucial for visualizing key performance indicators and data related to your software applications' performance, health, and availability. These widgets offer real-time insights into error rates, response times, and throughput metrics.

Create metric alerts directly from widgets in your Service and Database catalogs within the APM interface. These alerts function like standard metric alerts but are conveniently auto-populated and correlated with the underlying query from a specific widget. You can keep the base query or modify it according to your needs. Once an alert is set up, it appears in your Alert Management Screen with APM labels attached.

Manage triggered alert events from the Incidents card in the upper tier of your APM interface. The card displays if triggered events exist and their status. Select the status updates to view these events in your Incidents Screen with APM labels attached.

Auto-populated fields

The following fields are auto-populated when you set up an alert from the APM interface:

Field	Description	Example	Notes
PromQL query	Automatically generated based on the query underlying the widget and chosen data source. Dimensions and active filters defined during the alert creation is considered part of the query.	avg_over_time(sum(increase(duration_ms_sum{span_kind=~"SPAN_KIND_CONSUMER|SPAN_KIND_SERVER", service_name=~"my-app"}[900])) / sum(increase(duration_ms_count{span_kind=~"SPAN_KIND_CONSUMER|SPAN_KIND_SERVER", service_name=~"my-app"}[900]))[900]) * 1000	Query may be modified, aside from reference to service.
Labels	Labels will be used in Alerts or Incidents or to filter all alerts and incidents by service or widget name across services.	Service Catalog: product:apm-service-catalog, serviceName:<service name> Database Catalog: product:apm-db-catalog, apm-db-catalog.databaseName:<db.system>.<db.namespace>	Should not be modified. If a label is modified, there will be no correlation between an incident and its service.

Required permissions

Users with the following permissions can view and/or manage alerts.

Resource	Action	Description
metric.alerts	ReadConfig	View alert settings
metric.alerts	ReadConfig	View incident settings
metric.alerts	UpdateConfig	Create and modify alert settings

Learn more about roles and permissions here.

Create alerts

1. Navigate to APM, then Service Catalog or Databases in your Coralogix toolbar.

2. Select a service or database to view its Overview or Operation widgets.

3. Select the widget to use as the basis for your alert. Select the ellipsis and then Create Alert. If the widget has static metric categories (such as successes and errors), choose the category to be queried.

4. This opens a metric alert side modal that is auto-populated and correlated with the selected service, widget, and metric category selected. Define the remaining conditions and notification settings for a Metric alert or Anomaly Detection alert. See the use cases below for more information on the alert definition setup.

5. After defining alert and notification settings, select Create Alert in the upper-right-hand corner.

Use cases

Successful transactions & errors

This widget presents successful transactions and errors per minute over a selected time-picker interval. The interval produces aggregations every 60 seconds. In this case, you can create an alert for the service’s errors. A PromQL query is automatically generated based on the query underlying the widget and chosen data source.

The alert defined checks whether service <> has more than 100 errors for over 80% of 10 minutes. Every minute, the Coralogix alert mechanism looks at the last 10 minutes of query aggregations (produced every 60 seconds) to determine how many aggregations have breached the 100-value threshold. If this threshold is breached 8 out of 10 times, the alert is triggered and an incident is recorded.

Apdex score

The widget presents the Apdex score for a service every 60 seconds for 15 minutes. A PromQL query is automatically generated based on the query underlying the widget.

The alert defined will check whether the Apdex score of service <> is less than 70% for over 50% of 10 minutes. Every minute, the Coralogix alert mechanism will look at the last 10 minutes of query aggregations (produced every 60 seconds) to determine how many aggregations are less than the 0.7 threshold. If this threshold is breached more than 5 out of 10 times, the alert is triggered and an incident is recorded.

Manage incidents

View incidents by service

The Incidents column in the Service Catalog and Databases page displays incidents status for all existing services. The following indications are available:

• Red bell: The service has at least 1 triggered incident.

• Blue bell: The service has acknowledged or resolved incident(s) only.

• No indication: No alerts or incidents are configured for the service.

To inspect incidents:

1. Hover on the bell icon to display the number of triggered, acknowledged or resolved incidents per service.

2. Select the bell icon for the service to display the Incidents page with service-related incidents filtered by the service.

No triggered incidents

If you have set up alerts but no incidents have been triggered, the Incidents card displays this status.

Triggered incidents

If incidents exist for the alerts you’ve set up, the Incidents card will present the number of events that have been triggered and acknowledged.

Support

Need help?

Our customer success team is available 24/7 via in-app chat or email at support@coralogix.com to assist with setup and any questions.