I’ve considered setting our history service events cache to host-level and defining max size in bytes to make it more straightforward to manage memory levels.
But, I’m wondering if there would be tradeoffs here. Is there a risk of a lot of contention if there are many workflows running at the same time and there are a lot of shards on the host?
What considerations should we make when deciding whether to use per-shard or per-host events cache?
Thank you!