Deployment of AWS cluster with custom AMIs (vi – Monitoring the cluster)
AWS provides several tools to monitor and analyze usage of its many resources. The most widely used of these tools is CloudWatch, which covers much of AWS infrastructure. Although CloudWatch can be used to monitor several variables, HPC clusters are complex structures that benefit from more sophisticated monitoring tools. To accomplish this objective, AWS has incorporated several open-source Grafana and Prometheus add-ons. The said add-ons, which need to be invoked from the cluster configuration script, result in the availability of several visualization & instrumentation toolkits of interest to the cluster user. The CLUSTERs subdirectory details several examples of how to deploy these tools for some of the configurations previously introduced. The procedure to launch the cluster does not change and, once this is complete, the dashboards can be accessed via web browser at https://your_public_ip. The menu page will appear
Close Menu