“We stay in business through being very good at what we do and being very efficient with our time. Using tools such as Boundary helps us thrive and grow. Any time we make big changes to our network topology, I expect we will pay close attention to Boundary to verify the improvements that we expect to see, or spot any regressions.”
—Nick Zadrozny, Cofounder, OneMoreCloud
Through AWS monitoring tools, co-founder Nick Zadrozny began to keep a closer eye on billing and usage projections. In the fall of 2012, he saw that network usage suddenly began to grow.
“At the rate things were going, this increased usage was going to cost us an extra $5,000 a month or more,” he says. “I had wanted an excuse to try Boundary, and identifying rogue network usage seemed to be a perfect fit.”
Zadrozny added the Boundary service to the company’s configuration management, and within 20 minutes, Boundary was reporting from about 100 servers in an AWS region.
How Did Boundary Help?
After the rapid setup, Zadrozny was able to quickly analyze where the network was usage was spiking, helping him and his cofounder immediately focus their troubleshooting in the right place.
“Seeing per-second statistics from every process on every node helped us quickly identify one rogue process that was transmitting a few orders of magnitude more traffic than its peers,” he recalls.
Further investigation through Boundary indicated that the problem stemmed from a rare replication failure, which was continuously retransmitting data and wasting hundreds of gigabytes of network traffic.
After fixing the replication error, an administrator saw immediate confirmation in Boundary that traffic levels had returned to normal. Beyond helping the company avoid thousands of dollars in additional monthly usage fees from AWS, Boundary is saving hours of staff time in troubleshooting issues.
Now the company has a different perspective on system monitoring, based on real-time network statistics, and it has built alerts so they can prevent the replication error from affecting traffic or service levels. Over time, OneMoreCloud will use Boundary to create new alerts for significant issues so they can enhance the performance of the service for customers.
OneMoreCloud is an IaaS provider that hosts a search engine service for more than 1,000 customers, including those with small niche sites as well as sizable e-commerce operations. Through its open source software, OneMoreCloud enables customers to deliver search services, easily and without ongoing maintenance—while enjoying performance that is often significantly faster than internally sourced search functionality. The company runs the service primarily through the Amazon Web Services (AWS) public cloud, with additional private dedicated clusters on Rackspace and Linode.
…..learn about other Boundary customers at http://boundary.com/company/customers/
One Response to “OneMoreCloud: AWS Monitoring and saving hours of staff time in troubleshooting issues.”
Leave a Reply