Really easy given how much money it helps to save. What is challenging is to sell Pulumi to my coworkers
I have not heard of pulumi before, I'll check what is that. Our server bills are in upper hundreds and a k8s setup costs us almost the same amount. I think we'll have to optimize the resource requests in all the apps according to their more realistic usage to reduce the required compute/memory.
graph how much CPU, memory each service takes (avg, min, max, p95) and then set resource limits accordingly. Downside nodes after that
I will do that 👍
Обсуждают сегодня