I *think* that #TIL that #HAProxy's stick-table synchronisation between multiple peers is pretty flawed, if you're trying to use it for multi-node rate limiting: stats are *not* aggregated. To quote the manual:
> Each instance pushes its local updates and insertions to remote peers.
> The pushed values overwrite remote ones without aggregation.
I've not used HAP in a while (I'm getting started again!), but is this news to any other HAP admins? #SysAdmin #DevOps
@jonathanmatthews people usually use the stick table aggregator to achieve this, but this is an HAProxy Enterprise only component. https://www.haproxy.com/blog/introduction-to-haproxy-stick-tables
@wlallemand Sure, I'd found that component. Unfortunately it's not practical given that I can't see pricing information to find out if it's anywhere near to being a viable option. I'm just bootstrapping a small SaaS - thinking about relying on s/w whose acquisition process *starts* by requiring me to come cap in hand, asking for permission to be considered to try it out, before I'm privileged enough to be told how much I'll pay is very, very, *very* low down my priority list!
