What the HA module includes today
The current High Availability module provides a controlled active/passive data plane: continuous PostgreSQL streaming, minute-by-minute configuration, certificate and media synchronization, independent witness leases, fail-closed fencing and assisted signed role changes.
Backups and recovery remain part of the platform safety baseline. The HA module adds monitored multi-node operation and repeatable failure drills; it does not promise zero downtime or preservation of calls already in progress.
- Nodes & placement registry with explicit PRIMARY/STANDBY roles and fail-closed call admission.
- Fail-closed placement preflight and call-aware drain before maintenance or migration.
- Independent HTTPS, TLS, SIP, host and backup monitoring.
- Streaming database replication plus synchronized configuration, certificates, recordings and call captures.
- Signed witness lease, fencing and documented active/passive role reversal with recorded evidence.
Edition, deployment and price
High Availability is included in Enterprise. It can be purchased for Business, Contact Center and Service Provider for EUR 99 per month or EUR 990 per year. On relaypbx.pl the published gross price is PLN 549 per month or PLN 5,490 per year.
An HA add-on order starts with a free technical review. Checkout opens only after the deployment topology, DNS, nodes and replication approach have been accepted.
The module is available for Self-hosted and RelayPBX Dedicated Cloud only. RelayPBX Cloud redundancy is operated at the shared platform level and is not sold as a dedicated tenant topology.
- A second comparable node, storage, network addresses and provider charges are separate infrastructure costs.
- RelayPBX Dedicated Cloud and Self-hosted topology design require technical acceptance before an availability target or SLA can apply.
- Enterprise cannot be replaced by combining lower-edition modules; offline activation and the remaining Enterprise controls keep their own entitlement boundary.
Acceptance boundary and roadmap
Continuous state replication and controlled failover are implemented. Unattended failover and atomic traffic switching through Floating IP or an approved load balancer remain roadmap items until they pass host, network and dependency failure injection plus repeatable RPO/RTO acceptance.
- Record the topology, failure domains, quorum assumptions and owner of every shared dependency.
- Prove a clean restore and controlled role change while active calls are drained.
- Measure data loss and restoration time instead of assuming RPO or RTO.
- Repeat the acceptance after material platform, database, storage or routing changes.