Hi. I’m having an issue with an eventstore that grows way to quick.
The fundamental problem is of course bad design, it’s the first eventsourced application I’ve worked on, and we went totally overboard, thinking we could save everything, so there 's sa lot of intermittent state that really could be discarded. Problem is it’s in the same streams as critial data, so it’ll take some refactoring to solve this, since I don’t think you can set $maxAge on individual events?
I also introduced projections to optimize readmodels. Works great, but it’s doubled the size of the database. When actually stopping to think it’s pretty obvious that every linkTo takes as much space as a small event I guess?
Is there anything to do short term, besides setting maxCount on status and scavenging? Writing this I realize there isn’t, but just for the odd chance there’s something I’m overlooking…
(This isn’t big as in big data btw, but the old relational version of the system had 3-4gb sized databases, and now we’re at 70-80gb, and the ops at customers sites wasn’t really prepared for this, hence the issues)
/Peter