Allow servers to have backups in place in case the initial goes down, or even better automatically share the load across instances (with usrt approval of course)
god yes, or anything better than the current. Maybe something like community ‘groups’ where one can subscribe to them all with a single click, or jsut remove some to their tastes. Tbh add a couple layers of that and it’d be an insanely powerful way to group similar communities
Setting up a Lemmy server outside of the golden path of using the Ansible template is extremely difficult. I do this professionally and I couldn’t get federation working properly when running Lemmy on my Kubernetes instance.
Figuring out why federation is failing is very, very hard.
Lemmy requires a lot of resources to run. You need a VPS that’s at least $20/mo to work adequately under any load. Disk storage requirements for the DB are also rather high.
Lemmy 0.18.2 has some horrendous N+1 DB calls, e.g. one query per language (173 of them) when you create a new community. This hamstrings databases that are not colocated onto the same machine, e.g. neon.tech’s hosted pg db. I expect this will improve with time as the codebase matures, yet…
where are the technical criticisms?
Make account migrating possible
When federating, load past content
Allow servers to have backups in place in case the initial goes down, or even better automatically share the load across instances (with usrt approval of course)
add multi community aggregation to aggregate communities of the same topic on different instances.
god yes, or anything better than the current. Maybe something like community ‘groups’ where one can subscribe to them all with a single click, or jsut remove some to their tastes. Tbh add a couple layers of that and it’d be an insanely powerful way to group similar communities
Setting up a Lemmy server outside of the golden path of using the Ansible template is extremely difficult. I do this professionally and I couldn’t get federation working properly when running Lemmy on my Kubernetes instance.
Figuring out why federation is failing is very, very hard.
Lemmy requires a lot of resources to run. You need a VPS that’s at least $20/mo to work adequately under any load. Disk storage requirements for the DB are also rather high.
Lemmy 0.18.2 has some horrendous N+1 DB calls, e.g. one query per language (173 of them) when you create a new community. This hamstrings databases that are not colocated onto the same machine, e.g. neon.tech’s hosted pg db. I expect this will improve with time as the codebase matures, yet…
Instance administration tools are sorely lacking.