Transparency is the whole point of PulseWatch, so this post covers what's changing, what we ran before, and above all why we made the call. If you only read one section, make it the one on the reason.
The short version
Where we're coming from
Our infrastructure has changed shape a few times, and it helps to see the sequence in one place:
In other words, this isn't a leap into something unknown. Both pieces of the new setup are things we've already run, and we're simply putting them in the order that makes the most sense for a monitoring service.
The actual reason: Vercel's functions are too unstable
Here's the honest answer, without the corporate fog. Vercel's functions are just too unstable for us to rely on as the backbone of PulseWatch.
That matters more for us than for most products. PulseWatch exists to tell you, reliably, whether your services are up. If the thing doing the watching is itself unpredictable, you can no longer trust the data, and trustworthy data is the only thing we sell. A monitoring tool that can't be relied on is worse than no monitoring tool at all, because it gives you confidence you shouldn't have.
A normal website can survive an occasional hiccup in a function. A monitoring service can't, because a flaky check can look exactly like a real outage, or hide one. Stability isn't a nice-to-have for us, it's the product.
So what does that mean in practice?
- Vercel is no longer a primary dependency. We won't build the core of PulseWatch on something we can't count on.
- It still has a useful role. As a secondary, it adds a second, independent path without being the thing everything depends on.
- The primary needs to be boring. The US cluster is a setup we know well, and that's exactly the quality we want in the seat that carries the load.
Why the "US" cluster is in the EU
The cluster kept its original name, but it has moved. It's now located in the EU, so the primary workloads run in the EU as well. If you see "US cluster" in our notes or internal tooling, that's the one we mean. We know the naming is a bit confusing, and we'd rather explain it than pretend it isn't.
What this means for you
- No action needed. Your monitors and settings stay exactly as they are.
- A clearer failover story. A stable primary with Vercel behind it as the secondary gives us a second path if something goes wrong.
- More dependable checks. Moving the main load off the part we trust least is a direct win for the quality of the data you see.
- Same promise. Reliable, honest uptime data, with no surprises.
Infrastructure changes affect how much you can trust the data we show you. We'd rather explain a reshuffle, and admit that a platform didn't work out the way we hoped, than have you discover it by accident.
A word from our former CEO
I'll be direct: this was not the outcome I wanted. Vercel was a platform I believed in, and we gave it a fair shot. But when its functions proved too unstable to lean on, the decision became simple. PulseWatch is a promise about reliability, and I'm not willing to build that promise on something I can't rely on myself.
Bringing the US cluster back as primary, now hosted in the EU, is the responsible move. It's the setup I trust to carry the load, and keeping Vercel as the secondary means we still get value from it without betting the product on it.
Thank you for your patience with us as we keep refining things, and for holding us to the standard of telling you what's really going on. That's a standard I want us to keep.
Marco, CEO until May 2026
Comments
Questions, corrections or thoughts on this post? Join the thread below.