That's a good question. We've been doing this for ~year, whereas Chrome on App Engine/Functions has _just_ recently come out. So we've got a bit of a head-start and our infrastructure is pretty hardened at this point. Also a few other points:
- We've got some REST-based API's that make it dead-simple to interact with Chrome for the majority of the use-cases out there. Plus you're not restricted to having to maintain node-based infra if that's not your stack.
- A lot of good tooling has been developed, for instance a live-debugging tool that lets you visually see the browser (located here: https://chrome.browserless.io/).
- The nature of a lot of folks' business restricts them from using big cloud providers like Google, Amazon, and Microsoft. Just because one of them pivots into your market doesn't mean that they're going to squash you.
- Finally, I've written a driver + am actively involved in puppeteer's repos (plus others). So, in a way, when you sign up for an account you get support from me as well, which has a _ton_ of value if you've never done headless work.
Can vouch for this. Joel is a machine. Not sure how he’s able to do all of this single handed but Browserless is one of those niches where someone who legit listens to their customer does a much better job than a ~trillion dollar company.
It’s simple. This is browserless’s bread and butter. For GCP, it’s just another feature for someone to get a promo.
One feature that's not available or easy to use is audio from the chrome session. Browser page navigation works well in most places, sound not always. We've been having issues with pulseaudio s/w channels just going bad.
