Sematext Alerts User Guide
Receiving Alerts when your software is misbehaving or underperforming is crucial for every developer. Good alerting practices include:
- Alert Rules that detect Anomalies and notify when key performance indicators spike or drop
- Criteria that trigger Alerts for monitored resources
- Thresholds defined for Alerts when selected metrics reach certain severity levels
- Notification channels through which the responsible team will be notified about the incident
Sematext Alerts help you address underlying issues before they affect user experience. There are three types of Alerts.
- Threshold - Alerts that are based on classic thresholds. They are triggered when something crosses a pre-defined threshold.
- Anomaly - Alerts based on statistical anomaly detection. They are triggered when values suddenly change and deviate from the baseline.
- Heartbeat - Alerts triggered when something you are monitoring, like your servers, containers, or your applications, stops sending data to Sematext Cloud.
Threshold and Anomaly Alerts can be triggered for both Monitoring and Logs Apps, while Heartbeat Alerts are only available for Monitoring Apps.
Sematext lets you manage Alert Rules across your whole stack. You can use various notification channels to get alerted when critical issues occur so you can take action and resolve them. Our Alerts easily integrate with notification and chat software.
- Slack, and many more.
Default Alerts for a Logs App:
- Anomaly Alerts for Error and Warning search queries
Default Alerts for a Monitoring App:
- Metric Alert for Disk Usage
- Heartbeat Alert if the Agent stops sending data to the Monitoring App
- Metric Anomaly Alert for misbehaving Nodes
In this example which is an Elasticsearch Monitoring App, the Java usage threshold and Elasticsearch Node Anomalies are integration-specific default Alerts.
You can view all the default and custom Alerts on the Alert Rules page. This is where you can toggle, edit or delete any existing Alerts.
Sematext Alerts can cover both complicated alerting scenarios with multiple rules for both Logs and Monitoring, as well as Events. A basic Alert Rule with an email or Slack notification is enough to cover elementary alerting needs.
Alerting on Logs¶
To create an Alert you need to run a search query and press Save Query/Alert Rule.
A Save Query window will open with the option to Enable Alert. When it's toggled, the Alert Type picker and Condition will expand, and the Notifications and Schedule tabs will be enabled. You'll see:
- Alert Type, to choose either Threshold Alerts or Anomaly Alerts. For more information on alerting capabilities check alerts documentation pages.
- Condition, to choose when the Alert will be triggered.
- Notification hook details with a default email notification hook used to send the message to your account. You can also add additional email addresses as well as other types of notification hooks.
- Schedule, to choose when the Alerts will notify you based on a period, schedule or time.
Press Save and you're done. Check out the integrations documentation for more information about alert notification hooks.
Alerting on Metrics¶
Monitoring Apps have metric-based Alerting. Metrics have both Threshold and Anomaly Alerts which can be created on a per-metric basis.
Monitoring Apps also have Heartbeat Alerts as part of their settings. You can find it in the App Settings dropdown menu. They are triggered when what you are monitoring, like your servers, containers, or your applications stop sending data to Sematext.