Monitoring

This section should describe how to monitor applications running on the STUPS infrastructure.

CloudWatch Metrics

The most basic monitoring can be achieved by the out-of-the-box AWS CloudWatch metrics. CloudWatch monitoring is automatically enabled for EC2 instances deployed with Senza.

CloudWatch EC2 metrics contain the following information:

  • CPU Utilization
  • Network traffic
  • Disk throughput / operations per second - only for ephemeral storage, EBS volumes are not included

Taupage Monitoring Features

The Taupage AMI supports a few features for enhanced monitoring:

  • Enhanced CloudWatch metrics to monitor memory and diskspace: enable with enhanced_cloudwatch_metrics property in Taupage config (this allows monitoring RAM usage and root filesystem on EBS)
  • Prometheus Node Exporter to export system metrics: the Prometheus Node Exporter is automatically started on every Taupage EC2 instance on port 9100

ZMON

The ZMON Zalando monitoring tool can be deployed into each AWS account to allow cross-team monitoring and dashboards. Make sure that ZMON appliance is allowed by security groups to connect to port 9100 of monitored instances.

ZMON allows querying arbitrary CloudWatch metrics using the “cloudwatch()” check command.

ZMON allows parsing the Prometheus metrics using the the “http().prometheus()” check command.