SolarWinds Cloud Monitoring for DevOps

Let's say our support team is receiving multiple calls from end users complaining that one of our websites is slow.

Support asks the web team to investigate.

In the Pingdom dashboard, they see the site is up and available, but wait times are extremely high for one of the page elements.

Using Papertrail, the team notices a significant increase in retries and time-out errors.

A log velocity graph shows a spike in error volume.

One of our SREs is brought in to dig into the problem more deeply.

The SRE opens AppOptics, and sees that response time for the web tier is above the baseline. This increase corresponds to the rise in support calls we were seeing earlier.

By selecting some of the transaction outliers and performing a trace, we discover that the increase in response time is being caused by a high number of database requests.

The 'Host View' indicates that several MongoDB hosts are running hot.

Next, the SRE opens Loggly and looks at log data across the application.

There is a recurring error – too many threads are already waiting for a connection. This suggests that the system is running out of resources. By searching that error code, the SRE sees unusually high volume.

The "View Surrounding Events" tab shows that the rest of the system is working normally, so the problem appears to be a database, resource-allocation issue.

The SRE brings more database resources online to handle the increase in demand.

AppOptics and Pingdom confirm that the system is working correctly now that the new resources have been added.

The SRE notifies the web and support teams that the problem has been fixed.

Pingdom, AppOptics, Papertrail and Loggly combine to help DevOps professionals identify and troubleshoot problems quickly and easily.