Infrastructure experts Mesosphere have pushed their DC/OS Monitoring framework out of the preview stages, making v1.0 generally available to all.
The service is meant to help operations to stay on top of the status of DC/OS (Mesosphere’s cluster manager, container platform and operating system), its components, and services on a DC/OS cluster they are overseeing. It makes use of metrics-related tools such as Prometheus, Grafana, Telegraf and AlertManager for insight into a system’s health and performance.
To lower the entry barrier for those new to these components, DC/OS Monitoring includes a couple of preconfigured Grafana dashboards for visualisation which can be accessed via the DC/OS UI as well as preconfigured alerts. Ops with stable monitoring in place can integrate DC/OS Monitoring via Telegraf, which can also be used to export metrics of custom application to the service.
Compared to former versions, v1.0 allows the setting of pre-reserved roles for Prometheus and features a configuration option for the Grafana dashboard theme style. A DC/OS Defaults folder has been added to load DC/OS Grafana dashboards to, and Enterprise customers can use DC/OS CA certificates for verification in the Prometheus agent discovery talks.
DC/OS users who have already looked into the tool will be interested to learn that its CLI is now started with monitoring instead of the dcos-monitoring command and that Pushgateway is now switched off by default. The latter means that there’s no way of easily upgrading from v0.5, so you’d have to start from scratch if you’ve grown to like this monitoring option.
The DC/OS Monitoring package is available via the Mesosphere Catalog. To get going, you need to have DC/OS 1.12 or a newer version as well as the project’s CLI installed and be logged in as a superuser. Once deployed, the Monitoring service automatically discovers cluster nodes and sub-components. More information on the setup can be found in the “Getting started” guide.