It’s different.
Mastodon:
- Zero downtime software updates
- Can scale sidekiq workers and web front ends
- Leverages Elastic Search (optional) and Redis
- Terrible Performance, so the scaling is required
- More complex running and maintenance
- Hard coded character maximum (PRs to fix this have been denied)
Pleroma:
- Way faster, and much more performant
- Lots of more backend features
- Post size limit is configurable
- Has to be taken down to update (I don’t care if it’s “really fast,” it’s still a down-time release).
- Cannot scale at all. You can run one instance of Pleroma. That’s it. No support for multiple workers.
Both have a hard bottleneck of the Postgres database. Both use postgres and postgres doesn’t support multiple masters.