Recommended ECS alert conditions

New Relic's ECS integration reports and displays performance data from your Amazon ECS environment. This document provides some recommended alert conditions for monitoring ECS performance.

Recommended alert conditions

Here are some recommended ECS alert conditions. To add these alerts, go to the Alerts UI and add the following NRQL alert conditions to an existing or new alert policy:

  • High CPU usage

    • NRQL: FROM ContainerSample SELECT cpuUsed / cpuLimitCores

    • Critical: > 90% for 5 minutes

  • High memory usage

    • NRQL: FROM ContainerSample SELECT memoryUsageBytes / memorySizeLimitBytes

    • Critical: > 80% for 5 minutes

  • Restart count

    • NRQL: FROM ContainerSample SELECT max(restartCount) - min(restartCount)

    • Critical: > 5 for 5 minutes

For more help

If you need more help, check out these support and learning resources: