It’s no secret that Hyperic is a web-driven open source project and business. All our web properties from hyperic.com, to the wiki, forums and support are invaluable resources for us and our users. When we first started, we managed all these properties ourselves. In fact, as many scrappy startups do - they all started on the same box! Of course, we had an advantage - we used Hyperic HQ to do application performance monitoring to scale up our sites and make the most of the hardware we have. We also used it to assess capacity planning as we needed to scale out the system to meet our growing demand.

Earlier this year, we hit the limits of our existing hardware. We decided to move our wiki to its own setup and chose to move it to Contegix. Contegix of course, is a managed hosting provider that uses Hyperic HQ and specializes in managing Atlassian products like Confluence - Hyperic’s wiki. The migration and performance has been flawless, we know this since they share visibility with their customers to Hyperic HQ Enterprise. That is, until today.

I get an email this morning from Contegix Support:

Team:

We have recently received alerts for support.hyperic.com on hyperic01. At that time, the site was verified non-responsive. Upon investigation the instance had exceeded it’s allocated heap space. I’ve bumped this value from 1280 meg to 1536 meg, and restarted the instance. The site is responding properly now. Let us know if you need anything further.

Thanks,
-Chris

—
Contegix
Beyond Managed Hosting(r) for Your Enterprise

Since I was away from my computer, I check in with our Operations Admin first to find out how long it was going on and if he thinks everything is under control, and he says to me:

They were incredibly fast! We monitor the performance of that box with checks every minute. They responded and fixed it within that first minute we noticed it was having problems. Before I even had time to reach out to them, it was already fixed. That’s pretty much sub-minute response time, which just rocks.

It’s a nice reminder that eating your own dogfood has its benefits. One, it shows we have just as big of a stake in our software working as our customers, and two it has a nice by-product of getting to be your own customer testimonial! I think we should do more of it actually - I am going to organize some HyperCASTs with our scalability testing team, and ops teams to show some of our best practices in doing performance management and web operations for our own stuff. So stay tuned. More dogfood testimonials to come!