Initial thoughts
1. Are autoscaling runners better?
2. Advantages of a single server GitLab runner
3. Traditional cache tasks sequence
4. Searchi...
For further actions, you may consider blocking this person and/or reporting abuse
It depends on how your cook your k8s setup. Take a look at our new GitLab runner SaaS, it's beating anything in the world in cost/performance terms and it's built entirely on Kubernetes. And it's also running each job in the clean new VM, so the security is even better than on a single server.
puzl.cloud/products/ci-runners-for...
Interesting, what do you do differently than most people ?
A lot of stuff really: from host preparation, fine-tuned container runtime and filesystem to our unique billing model. You pay only for cpu-seconds and memory GB-seconds consumed (not just allocated) by each pipeline job. We call it Burstable Resources, and such billing concept isn't available on any cloud platform. So if you just take a normal k8s GitLab runner and run it in some GKE cluster, you won't be able to achieve such cost/performance profile even with tuned autoscaler.