Ever since Hurricane Sandy, which brought down our service along with many others, we have been thinking of ways to further improve our service reliability, and increase our protection against catastrophic events. In the past few month, we’ve also encountered downtimes stemming from our web hoster, which further strengthened our resolve to improve our hosting situation.
Thus, as part of the newly-minted measures designed to improve our service, we will be migrating our servers on 20 Jan 2013 (Sunday) from 4am to 8am GMT. This operation is being intricately planned to minimize downtime. Conservatively, we are setting aside a four-hour maintenance window. Realistically, we believe this operation can be performed with a much shorter disruption to our service.
During this period, you might not be able to log into the dashboard / chat via IM. This downtime will not affect the loading of your website, so no worries, you need not make any code changes at all.
Read on if you love details..
“How does this improve Zopim’s reliability”, you ask..
Private server rack
Over the past weeks, we have been setting up our new servers. Our hosting partners have kindly agreed to place these servers in our private server rack. Usually, we Zops love sharing, but this is one instance where being isolated works great. The fact we’re on a private network means better connectivity between our servers, and a lesser likelihood of being affected by other users on the network. Also, there are plenty of empty slots on the rack, which allows us to easily scale up our servers in the near future.
“Pics or it didn’t happen!”
Higher level of support
We now receive top-tier support from our hosting provider, which includes direct & instant access to highly skilled server admins 24 x 7, allowing us to bypass the lower level support staff. This translates to faster resolution times, should our servers run into problems.
Earlier, with most of our servers in one location, and having only one hosting provider, we had a single point of failure. We have since added redundancy by replicating our servers on the cloud, and engaging more providers.
The global footprint of these providers also allows us to eventually add servers in multiple locations worldwide, further improving redundancy and performance.