This page is not created by, affiliated with, or supported by Slack Technologies, Inc.
2021-05-29
Channels
- # announcements (3)
- # aws (12)
- # babashka (11)
- # beginners (46)
- # calva (10)
- # cider (6)
- # clara (3)
- # cljdoc (6)
- # cljs-dev (13)
- # cljsrn (2)
- # clojure (49)
- # clojure-europe (2)
- # clojure-gamedev (1)
- # clojure-germany (22)
- # clojure-uk (2)
- # clojurescript (28)
- # clojureverse-ops (8)
- # conjure (6)
- # cursive (2)
- # emacs (1)
- # figwheel-main (9)
- # heroku (12)
- # jobs-discuss (1)
- # malli (10)
- # off-topic (1)
- # practicalli (8)
- # re-frame (25)
- # reagent (6)
- # shadow-cljs (24)
- # testing (10)
- # vscode (4)
Has anyone migrated from AWS EC2 + Kubernetes to AWS Fargate + Kubernetes? Any pros and cons of doing this? The goal is to reduce costs.
Best to check pricing but EC2 is always going to be cheaper than Fargate, but it involves more management and prep. We're running all workloads on ECS+Fargate, but not because of the direct cost, but because it cuts human cost required in managing everything.
One potential con, depending on your workload, is that Fargate has more restrictive cpu/memory limits. So if you need more than 8 vCPUs per container, Fargate may be a bad fit.
It’s be great to have 8. When you said that I got excited since we’d love more vcpu per node.
I've recently run into this... but yes - refactoring code and fixing hotspots quickly addressed that
Thanks for the input, all. I think I’ll let the Fargate idea pass since we have been running with a decent setup of EKS + EC2 + Chef + Terraform and the cost savings might only be marginal if any.