Maintenance over the next few weeks


In the last few months, we've seen Helpmonks growing a lot. The Helpmonks growth means that we are in need to extend our infrastructure. While we already deploy the best practices for an always-on service, we think we can do more.

After a thorough examination of all our options for a hosting partner, we’ve decided to move our entire infrastructure, except the database, to the Google Cloud Platform. Google Cloud Platform allows us to dynamically add new servers if there is a peak in requests. But far more important is that we can extend all Helpmonks servers to make them available closer to the location where you sign in. Serving Helpmonks closer means better performance for you.

As mentioned above, we will also host our MongoDB database with another partner. For that, we’ve decided to move the database to MongoDB and their hosting service. Their hosting environment has a stellar uptime reputation, and the database is automatically replicated over three different servers.

That said, in the next few weeks, we will be undergoing a lot of migration work. Though most of the migration will be behind the scenes, i.e., without any downtime for you, the change to the database will cause downtime.

Therefore, Helpmonks will not be available this Sunday, March 24th from 10 am to 2 pm EST (Eastern Standard Time). Incoming emails are not affected during the migration and will be processed when Helpmonks is back online.

You can follow us on Twitter or check our status page to get real-time updates for the maintenance. We will do our utmost to shorten the downtime and think that we will be back online sooner.

If you have any questions, please feel free to reach out to us.

Nothing against MongoDB and their dbase hosting service, but I think you will discover that hosting your dbase closer to your app servers (i.e., within the same subnet) will be required for the performance that you seek. Regardless of how fast both hosting services are individually, I’ve never had good luck separating dbase and app layers within separate data centers. Even within the same data center, if it is too many hops to the dbase, you may experience performance problems that don’t even flag or create errors very well… you will have issues you can’t pin down and will spend weeks trying to fix… and they won’t magically go away until you move them closer. Of course, I’m talking about 65,000+ active users-- I’m not sure it will matter in a smaller environment with lesser requirements.

Just my 2 cents–


1 Like

Not to brag, but I think I predicted your outages and performance issues?

Glad you got them resolved, there’s nothing like being on the same subnet–



No, not really. As outlined at “How we solved outages and performance issues in Helpmonks” we got the database up to speed but we’ve had to throw a lot of process power at it.

Apparently, Amazon, Google Cloud, etc., all throttle their disks by I/O. Depending on the machine you buy you also get certain I/O performance with it. However, it stops at a certain point. Also, you will need to upgrade to “premium Ethernet” connection.

In the end, you just keep on paying “premium” for just about everything. While we did all this, and then still saw that it’s not any better than running it ourselves, we reverted the operation.