Here at Sencha, we’re constantly aware of how much you depend on our frameworks, tools, services and forums — to learn new skills, to develop amazing apps, and even just to hang out with our amazing community.
But we also know how frustrating it can be when things that you rely on on a daily basis suddenly aren’t, well, reliable. And how frustrating it can be when you don’t know why.
So, we’d like to apologize for some of the issues we’ve been having recently with our web server, and in particular, the performance of our forums, documentation, and main site. And explain how we got there and what we’re doing to ensure that it doesn’t happen again.
Firstly, like many other companies, we use Amazon Web Services to provide much of our web server infrastructure, and were affected by that service’s downtime several weeks ago. That certainly caused us to think carefully about how we wanted to provide more tolerance for our main site. Shortly afterwards, we launched Ext JS 4.0, a major upgrade to our flagship desktop framework, and accordingly we also doubled the number of our server instances to handle the demand.
Meanwhile, as part of the Ext JS SDK, we wanted to demonstrate exciting new features such as infinite scrolling and the buffered grid — and we needed to show their use against a real-world data source. Quite reasonably (we thought), we chose our own community forum’s database for many of our examples — a large and dynamic data source that we could easily create a server-side API for.
At this point, the unfortunate consequence probably writes itself. A highly popular SDK, containing a large number of highly data-driven demos, being used by tens of thousands of developers around the world all at once, via a larger pool of servers… all accessing a single database instance.
Sadly the database simply couldn’t take it in its stride. Although designed for the vBulletin forum software itself, it just didn’t have the appropriate optimizations for the sheer number of direct data queries that were being made.
So as a result, you’ll have noticed that forums have been running significantly slower than usual. And worse, on a small number of occasions recently, the impact has been significant enough to affect the other sites that rely on the same database server: namely the Sencha web site itself, our online documentation, and SenchaDevs.
But now that we have identified the cause of the issue, we’ve addressed it in a number of ways. We’ve optimized the database for the queries that the SDK grid examples are making, and for future versions of these examples, we’ll be using alternative data sources. We’ll also be distributing the database instances for different sites across different physical servers.
Finally, we’re installing European instances of our services so that developers to the east of GMT should have greatly improved responsiveness for all of these sites.
(As an aside, we’d also like to point out that our newly announced Sencha.io cloud services run on entirely different infrastructure. They are also proving extremely popular, but are designed to be completely scalable with demand, and have not been affected at all by this issue.)
So we’re on top of it, and are sure we’re able to put these disappointing events behind us. Our apologies again for any inconvenience caused.
And of course we hope to see you all on those blazingly-fast forums again!