When comparing Google Cloud Platform vs Gitlabs integrated CI, the Slant community recommends Google Cloud Platform for most people. In the question“What are the best continuous deployment Kubernetes platforms?” Google Cloud Platform is ranked 1st while Gitlabs integrated CI is ranked 7th. The most important reason people chose Google Cloud Platform is:
Tensor Processing Units are bare-metal units designed to train your Machine Learning models faster. This is exclusive only to Google Cloud Platform.
Specs
Ranked in these QuestionsQuestion Ranking
Pros
Pro TPUs
Tensor Processing Units are bare-metal units designed to train your Machine Learning models faster. This is exclusive only to Google Cloud Platform.
Pro Free tier
GCP gives your 365 days of free trial and 300 US$ credit including a Always Free if you upgrade.
Pro Live-migrating VMs
Your VMs will still run even if there is a maintenance work going on to the tenant running your VM - thanks to Google's Live-migration system.
Pro Clear IAM roles
Unlike most Enterprise-grade IAM role systems, GCP has a clearer IAM permissions model, with Azure following.
Pro First-class Kubernetes support
Kubernetes is a project from Google's Borg. GCP is the first Kubernetes-Certified infrastructure.
Pro Secure and efficient to the core
Google datacenters are built with minimal security issues in mind while still reducing carbon footprint, making Google datacenters a much more greener infrastructure than competitors.
Pro Runs on the same infrastructure with Google's Services
Google's datacenters runs your favorite video streaming services and even Google itself. Discord also relies on GCP to handle its workloads.
Pro CI is integrated into the version control system
Pro Open-source
Pro Easiest CI for Windows binary, because of Docker-Windows Runner Executor
Cons
Con Expensive
GCP might be cheaper than AWS but it is more expensive than Azure. The main reason for this might be because of the extensive investment done for datacenters being efficient and green and secure.
Con Live Migrating VMs does not apply to GPU-attached VMs
Because this kind of VMs has a direct passthrough to the Processing Unit, this is not available and restart only will be available for the VM attached.
Con Unstable Egress/Ingress
Google Cloud Platform is a bit unstable with its egress and ingress. You may receive subtle blips when maintenance work is being carried on the routing.
Con Security risks
Con Only works with gitlab
Con Not lightweight
Not a lightweight solution, demanding and memory hungry.
Con No Windows support
No Windows support, but it's possible to use a Bitnami stack.
