Relay sizing in AWS

If y’all have relays in AWS, what sorts of sizing configs and OSes are you using?

Most of my relays are currently CentOS/x86_64, so I presume I’d either use exactly that, or Amazon Linux2 on an AMD processor.

AWS or any other compute provider, size can be reviewed here based on child client count: https://bigfix-mark.github.io/Docs/BigFix%20Capacity%20Planning%20v10.pdf

Graviton would be less expensive then x86_64 but Relays are not (yet) supported on Graviton. AL2 is not officially supported for Relays but it does work for us. https://support.hcltechsw.com/csm?id=kb_article&sysparm_article=KB0087327#linux

In addition to the bitness and distro, I was thinking about the configs. Seems like t3a.small or t3a.medium match relays pretty well. (Unless the t series isn’t appropriate for the usecase?)

(I’ll soon file an idea on supporting Relays on AL2/ARM. It’s to the customer’s economic advantage.)