You're trusting DotNest with your Orchard sites because you want some convenient, flexible, and reliable hosting. It's surely convenient (just a couple of clicks and you have an Orchard website running), and it brings all the flexibility of Orchard Core and Orchard 1. However, about that reliable part: How reliable is DotNest, actually? Pretty damn reliable! So, let's look behind the scenes and see some numbers.
When it comes to the availability of a web app, we're talking about it serving users (requests) with correct results (i.e. not just error pages but what they're actually looking for) within some performance metrics (like if a page takes a minute to load then technically it's working but we consider the app unavailable). If it does this almost every time, then we can say it's reliable. What does this look like for DotNest?
You may already know that DotNest runs on Microsoft's cloud, Azure. We collect telemetry data about how the platform performs with the Azure feature Application Insights: It's quite a useful tool to see all kinds of metrics and charts about how fast DotNest is, whether there are any errors, and all kinds of details down to the individual requests. It also immediately notifies us if something is off, even if you don't notice anything wrong with your site! But for cases when you do notice, we also use Application Insights availability tests: This pings endpoints of the platform periodically, much like an ordinary user would, and verifies that they load correctly and within 30 seconds (this is the minimum threshold that's possible to configure). It does this from multiple locations around the world, so we can also know if there are network issues.
For DotNest, Azure pings the app every 5 minutes from 5 locations. A failure means that the request's response is incorrect or takes more than 30 seconds. If it fails once, then it'll try again within 20 seconds, and then once more again; the app is considered unavailable if it still fails on the third try. Here are the availability metrics of one of those endpoints for the Orchard Core sites for the last 3 months:
Yeah, that's 17 failing availability checks in 3 months! The tiny blip you may notice on the chart around mid-July was because of that; this was the only outage we had, for a total of 5 minutes. Sounds pretty reliable, right? All this translates to an availability of 99.99% which is something we're quite proud of being able to provide for you!
BTW the reason why checks from the East US region are the fastest is because the app is hosted in the East US data center of Azure. For the rest of the locations, it takes this much for the data to travel back and forth via the internet.