Metrics & Log Alerts: Anomalies, Thresholds, and Heartbeats
The event stream captures all your key IT and operations events. By default, all alerts are streamed as events. A single look at the event stream gives a good idea of the current health of your whole infrastructure and your applications:
- Create an alert rule on any metric or any log stream
- Use metric filters to narrow down alert rules (e.g., alert when any host tagged with “type:elasticsearch” and “role:datanode” has cpu > 80%), requiring no reconfiguration as new nodes join and are tagged
- Choose between threshold-based alerts and anomaly detection underpinned by a number of statistical algorithms trained to detect when data deviates from the norm
- Get notified when application fails to report metrics or logs within a given time frame, set by you, giving you full control over notification speed and sensitivity
- Turn any log search into an alert query (e.g., alert when count of logs matching “severity:ERROR or exception” > 100/min)
Alert Notification Hooks
You can ship your Sematext alerts wherever you need to:
- Use existing integrations with Slack, PagerDuty, VictorOps, OpsGenie, etc.
- Have a custom WebHook you’d like to call? No problem – define the endpoint, format (JSON, HTTP params, form-encoded), optional URL parameters and optional HTTP request headers.
- Want alerts via email? That’s the default, of course.
Email notifications include embedded charts, so you can quickly assess the situation.
- Like to get your alert notifications via your mobile device?
Use Pushover integration!
- Still using Nagios? Yes, you can get your Sematext alerts sent to Nagios.
- Need to ship your Sematext alerts somewhere else?
Use Zapier integration.