Forget - performance question

I’m having some problems with restic, but now accumulated repos with >1000 snapshots. I’d like to forget many of them, but also read that forgetting is a slow process. Is it more advisable to frequently forget small pieces of data, or should I aim at forgetting many pieces of data at once, but only every now and then?

There’s nothing wrong with running forget with a policy regularly. Even running it after each backup run isn’t uncommon. Note that prune is a separate step to clean out unused data in your repository.

forget itself is pretty fast, it’s the prune part that takes more time. However, recent restic versions can often prune a repository within minutes (although that obviously depends on the repository size).