a
apetkovic
I'm looking for the current status and roadmap position of multi-zone high availability for Web Services.
Our SOC 2 Type 2 controls require production service instances to be deployed across physically separate data centers. The requirement would be satisfied by something close to what Render already ships for Postgres HA: instances in the same region, placed in zones that are geographically separate, with automatic failover when a zone becomes unavailable. We do not need multi-region.
My understanding from the public docs is that Web Service instances carry no zone-diversity guarantee, even when a service is scaled to multiple instances. If that is wrong, I would like to understand the actual placement behavior.
Three questions:
  1. Is zone-redundant placement for Web Services on the roadmap? If so, is there a rough timeline or a way to follow it?
  2. For a Web Service scaled to multiple instances today, is there any guarantee, or documented best-effort behavior, around spreading those instances across zones?
  3. Is there documentation on Web Service instance placement and failure domains that we can reference as audit evidence?
This is currently blocking us from running Web Services in production. We would prefer to keep this workload on Render rather than move it, so any clarity you can give helps.