Using Sensu for Alerting and Historical Analysis

In this testimonial, DevOps Manager James Sandlin at eBay discusses previous monitoring solutions (including Nagios, Icinga, and Zabbix) and why they’re using Sensu Go. Instead of the “old-school way” of setting up things — which doesn’t work with dynamic scaling — he likes that Sensu enables him to start with automation from the ground up. With Sensu’s ability for de-registration, he knows he won’t get a deluge of alerts when he deletes a bunch of build nodes. He’s also a fan of the Sensu Community and is looking to use Sensu as both an alerting and historical analysis tool — providing uptime metrics and historical reporting as well as real-time status for customers. Already he’s found that Sensu has alerted the team to issues that they wouldn’t have caught until users “come screaming” — watch the video to learn more.

The problem with the old school monitoring tools

eBay had been working on getting the automation in place for close to three months and had tried Nagios, Icinga, and Zabbix.

Most monitoring tools are based upon the old school way of doing things: you set up a server, you install software, and it remains in that state for many years in that static configuration. And if that server goes away, you’re going to get alerts, which works great until you move into a dynamic scaling IT environment.

The problem with the old school monitoring tools in a dynamic cloud environment is when a few hundred compute nodes are deleted intentionally, the central provisioning cannot keep up and now you have a few hundred servers sending false alerts. You don’t know what is and what is not a real alert, which makes it difficult to do any troubleshooting.

Why Sensu

With Sensu’s auto registration/de-registration, James knows they’re not going to get alarm floods when they delete a number of compute nodes.

The integration with other best of breed products, including various time-series databases and native handling of metrics, allowed them to use their existing solution for historical analysis and reporting. The Grafana integration allows a flexible UI that any developer can easily use with Sensu’s data collection. The active community and a large number of plugins are additional benefits of Sensu’s OSS model.