Great tool with some caveats

Google Compute Engine usually delivers good and predictable performance for our self-hosted stack of applications. However, when running n8n heavy workflows, even with a tunned instance configuration in docker, performance usually peaks usually due to memory usage. We often need to upgrade memory when we have heavier data processing and workflows. Other than that, in a normal usage, we don't really have any problems

I couldn't give a perfect score here because this is one of our pain points with GCE. We occasionally experience some of our instances becoming unresponsive or not responding to health checks. These are somewhat rare incidents, but when they happen, causes disruption for our company's employees and I believe this would be a serious issue in production at scale in a external application.

I think Google cloud is well suited for both startups, medium enterprises and big corps, because of its pay-per-use pricing model, with a generous initials free credits program. It ranges from both simple to heavy users, with a robust solution for the entire management of all parts of a VM, including disks, OS, and the hardware itself, allowing to scale the size with ease. The only downside I see, it's because, sometimes, our users experience some error 500 due to the machine restarting without notice.