Freek Van der Herten, Co-founder of Oh Dear, on Laravel, Monitoring, and AI | E133

However, there was a Cloudflare outage and you were not happy about it.

You want to jump on that and discuss that?

We actually, in our history, had quite a few of Cloudflare outages.

I don't know which one you are referring to.

I saw it on

your blog somewhere and you were not happy about it.

Yeah.

You know, if you're like an uptime service like ours, And a big outage happens, this has a big effect on your load.

Because if a website is operating okay, what happens is we check it every minute and everything going along.

It's online, okay, we don't have to do anything.

But if a website is down, we actually have to do more work because we are going to check it faster.

We're going to check it half a minute or maybe even faster just to... to just keep an eye on it, we have to send notifications out.

So we send out mails and Slack alerts and whatnot.

And so if something like Cloudflare goes down and basically uh three quarters of our users their website goes down yeah we need to do a lot more work uh because we're operating a bit at scale and yeah then uh it's it's especially mathias who has like uh a little bit of a harder time because he keeps the track on the load of stuff and we really see on our load like there's something happening there we actually have like internal alerts for this like if the load goes up in a spike or if we send out more notifications than usual then we already know there's something up with the internet because we need to do a lot more at this point so when something like that happens a big outage yeah that's that's all always a little bit of an exciting moment for us uh as well

gotcha gotcha and you know cloud fair now powers i don't know half the internet depends on it i don't know something like that so i heard this word infrastructure concentration risk okay so considering what you do you know and consider so cloud fair is one thing but there could be other things also of this word infrastructure concentration risk okay how have you guys under the hood sorted this out because so many customers so many websites tens of thousands millions maybe are depending on oh dear

yeah

so how

do you manage that how can how can customers all over the world you know attach websites and be you know sure that you guys have taken care of them

It's more of a question for Matthias for the bigger underlying points.

But we do have some safeguards in place.

So, Odeer, it's one application, right? But we actually have multiple servers that check the uptime of a site.

And those checkers, they're all around the world, but they are also from multiple vendors.

so we always make sure that if we send out an alert we've checked it with multiple vendors we checked it with multiple places around the world so we're absolutely sure that we're not going to sense any false positives because one one party is down and that's that's basically how we handle that We're also resilient in a way that if our main server would go online, we have a second backup on another location so we can quickly switch to another.

So these are the kind of scenarios that I hope we will never have, but we're prepared for because we're aware that a lot of people rely on us.

And yeah, if the uptime checker itself goes down, that really isn't...