At a glance
Read Replicas is the current count of read-only Postgres replicas provisioned for your Supabase project. Replicas serve read traffic from a copy of the primary, taking load off the writer and putting data physically closer to your users. This is a topology read, not a performance read: it tells you how much horizontal read capacity is in place right now.A change in this number is almost always deliberate (someone scaled up for a launch, or removed a replica to save cost), so an unexpected drop is worth a glance. Zero replicas is normal for smaller projects; all reads simply go to the primary.
What it tracks
The card reports the count of Read Replicas for the selected period, sampled in real time (RT). Read replicas are a paid Supabase feature available on the Pro plan and above, provisioned per region from the project’s Infrastructure settings. There is no alert threshold on this card (-); the count is informational and changes only when capacity is added or removed. It is most useful alongside Read Replica Lag (seconds), which tells you whether the replicas you have are actually keeping up. Roles: engineering, operations.
Reconciling against the source
Confirm the count in the Supabase project dashboard under Database, then Read Replicas, or via the Supabase Management API project endpoint, which lists each replica with its region and status. At the Postgres level you can confirm streaming replicas on the primary withSELECT client_addr, state, sync_state FROM pg_stat_replication, which returns one row per connected replica.