Clojurians Log v2

once it's set up, either one will be the same as far as usage goes. Setting up postgres will take slightly more work since you have to set up a postgres DB, but not a big deal. The main difference is backups. with postgres you can potentially use a managed service like digitalocean's which will handle backups for you. a downside is that you have to pay more for managed postgres. If you're ok with e.g. weekly backups, you could use rocksdb instead of postgres, and enable the automatic filesystem backups in digitalocean (or set up rsync cron job from somewhere). That's what I've been for my own apps since I decided I'm too cheap to pay the $15/month for managed postgres ha ha. With XTDB2 I'm going to look into a few options for kafka alternatives. e.g. there was one that can use just S3 as a backend. If that works it could be awesome. digitalocean S3 is only $5/month, then you could use that both for XTDB and any other object storage you need. I'll probably start working on migrating biff to XTDB2 in october.

I was concerned from some of the things I have seen that seemed to indicate that XTDB2 required kafka which seems like a huge headache to me. S3 would be an improvement

Yeah--there is an old discussion here where I asked about this: https://discuss.xtdb.com/t/deployment-recommendations-for-small-apps/483 basically there are a couple kafka-compatible things that could potentially be deployed on the same machine as your application.

Another one (kafka-compatible-on-s3 thing) I saw recently https://github.com/AutoMQ/automq

ooh, that looks interesting

Or just continue with XTDB v1 - it seems stable - and maybe (suggestion for @taylor.jeremydavid) Juxt could maintain and evolve it it as an XTDB Junior for projects like Biff

I'll definitely be migrating Biff to v2 🙂 but certainly no harm in keeping a project on v1