Skip to main content
Version: 3.1

Monitoring your cluster

Deprecation notice

Support for Portworx on the following orchestrators will be deprecated:

  • Docker Swarm and Docker Standalone
  • Nomad

For installation instructions on other platforms, refer to the Platform guides section.

note

This document presents the non-Kubernetes method of monitoring a Portworx cluster. Please refer to the Monitoring page if you are running Portworx on Kubernetes.

Using node_exporter and cadvisor alongside Portworx​

Node exporter and cadvisor with prometheus are great tools that export metrics about hardware/OS and containers. In order to provide metrics from the host, both tools require '/' to be exported into the container as /rootfs.

Portworx exports mounts to other containers and these in turn also get exported to node_exporter and cadvisor. In order for all mount events to be propagated to these containers, the root fs on the host should be bind mounted as "ro:slave". If this is not done, it is possible that these containers hold on to these volume mounts preventing the Portworx volumes from being used on other hosts.

# Host root filesystem should be mounted as read-only:slave
-v "/:/rootfs:ro" \

Monitoring Portworx​

Portworx exports metrics to Prometheus as are described here

Alert manager integration is described here

Grafana templates are provided here

Datadog integration with Portworx is described here

Last updated on