GCP is one of the most unreliable cloud solutions from a so called reliable provider - Google. In fact, so many services are a bad reflection of their lack of technical ability which operates like a third-rate product initiative. The entire platform is not only rigid but lacks sufficient variety of service offerings to match the multitude of domain applications. In fact, it cannot be stressed enough that the entire platform runs like an experiment where the services are at best buggy and over-priced. For any architectural decision, it will behoove you to wonder that it can most likely be done better on AWS and with a more flexible pricing option. In fact, there are no shortage of engineers and architects that are sufficiently qualified on AWS in the market to help an organization in their ramp-up time. GCP in many respects also lacks sufficient compliance and governance features. And, sudden robotic alerts are common to reek untimely havoc into your otherwise smooth running systems causing abrupt shutdowns, no advanced warnings, and extremely short three day violation messages. Online support is virtually none and trying to get hold of someone is extremely difficult - lack of human element is not only bizarre it spells unreliable alarm bells during critical outage situations. Most of the AI services, especially natural language related, are at best terribly executed and accuracy is atrocious. Everything on GCP is a bad reflection of what internally Google is really like: arrogant, disorganized, and overly bureaucratic where they forget about the customer's needs. In fact, even the monitoring service is terrible. Most of their databases focus on SQL, aren't we in a world where NoSQL is the norm? Google has always been terrible at understanding semantic graph concepts, literally everything they touch turns into a probabilistic problem of approximations. Where it lacks in quality and variety of services, it makes up for in devops innovations. It may be a good option where applications are still experimental and in the development stage. However, the reliability is so bad that it leaves little reason to build anything on the platform that may one day be for production use. By the time application is ready to go live, the ongoing frustration will drive one nuts on the platform to want to switch to AWS.
