Google Kubernetes Engine: Kubernetes clusters offer an easy-to-start experience and good integration with other Google Cloud services
January 22, 2022
Google Kubernetes Engine: Kubernetes clusters offer an easy-to-start experience and good integration with other Google Cloud services
Score 8 out of 10
Vetted Review
Verified User
Overall Satisfaction with Google Kubernetes Engine
Two products I work on are using Google Kubernetes Engine clusters. For the most part, the development efforts mostly go as far as "put service in container," so stuff such as scalability is left to 3rd party components that we use. The Google Kubernetes Engine can use a specific Google-provided ingress controller that is very beneficial when it comes to integrating with other services/products such as Cloud Armor, but it's also vendor-specific, so it has its own quirks and learning curve. Thus, we use the Google Kubernetes Engine just like a regular managed Kubernetes cloud service. The products we have in the Google Kubernetes Engine cluster deal with data piping, collection, and even some machine learning. The major problem that the Google Kubernetes Engine solves for us is a completely managed cloud Kubernetes service - we have an easier time managing our clusters (updates, scaling, and uptime SLA), doing physical and virtual migrations (moving nodes geographically, data in volumes, etc.).
- Engine upgrade rollout strategy - well documented and configurable
- Integration with other Google Cloud services like the Compute Engine, SaaS databases, and some cloud networking like Cloud Armor
- Graphical interface for a lot of operations - either for a quick peek/overview or actual work done by administrators and/or developers (via the Google Cloud Console, for example)
- It cannot reach true zero scale - they have a competing(?) product for that - Cloud Run Kubernetes clusters. It seems like the Google Kubernetes Engine may not be as flexible as some people need - in terms of costs and infrastructure.
- Some networking for the Google Kubernetes Engine is way too "hidden" from other similar services from Google Cloud - like network whitelisting (for the control plane), external IPs(s) are not a part of the VPC network overview, data storage.
- We had to make a hack for node-specific changes (max open file descriptors) because we put Elastic in our Google Kubernetes Engine clusters. These changes were made as hacks because there is still no official API/command approach to have such a form of control over the cluster's infrastructure.
- Ease of use and being newbie-friendly - several teams had to start using Google Kubernetes Engine without any prior K8s knowledge, so they managed to learn and apply on the fly.
- Integration with other Google Cloud services - SaaS database, networking, etc.
- Ability to use multiple geographical locations - migrations across locations or cross-zone/region clusters.
- Positive: Allowed us to start and produce working software regardless of our experience level.
- Positive: Integration with other Google Cloud services that we wanted to use anyway.
- Negative: It's kinda clunky, and some scenarios seem Google Kubernetes Engine-specific instead of being more integrated with other Google Cloud services or the web UI.
We had to move several products to Google Cloud, and the Google Kubernetes Engine was the option recommended to us, so we investigated it and ran with it. Back then (2019), we were not aware of Cloud Run-provisioned K8s clusters, so our other option was a completely self-managed K8s cluster on Compute Engine VMs, which we did not have the knowledge of and capacity to handle.
Do you think Google Kubernetes Engine delivers good value for the price?
Yes
Are you happy with Google Kubernetes Engine's feature set?
Yes
Did Google Kubernetes Engine live up to sales and marketing promises?
Yes
Did implementation of Google Kubernetes Engine go as expected?
Yes
Would you buy Google Kubernetes Engine again?
Yes