What's changing: PowerSync Service adds a configurable heartbeat_interval_seconds option for MongoDB, Postgres, and SQL Server replication sources. For MongoDB sources specifically, the default behavior also changes: if there's no recent resume-token update (typically due to replication lag from bulk writes elsewhere in the cluster), the service now writes a heartbeat to _powersync_checkpoints so the change stream can make progress once it catches up. This is aimed at avoiding the PSYNC_S1345 change stream timeout error, which previously could require re-replication from scratch to recover.
High priority — existing docs are now inaccurate or incomplete
- Error Codes Reference (
https://docs.powersync.com/debugging/error-codes) — the PSYNC_S1345 entry describes the timeout ("Failed to read MongoDB Change Stream due to a timeout... If this is not resolved after retries, replication may need to be restarted from scratch") but doesn't mention the new heartbeat_interval_seconds option or the MongoDB default-behavior change specifically designed to reduce/avoid this error.
For reviewer consideration — may be worth documenting
- Self-Hosted Instance Configuration (
https://docs.powersync.com/configuration/powersync-service/self-hosted-instances) — the example service.yaml replication config skeleton doesn't list heartbeat_interval_seconds as a supported replication.connections option. Candidate for addition if the team wants it documented as an advanced/tunable setting; the PR notes the configurable range may still expand over time.
Source PR: powersync-ja/powersync-service#747 — merged 2026-08-05.
Note: docs.powersync.com (llms.txt/llms-full.txt) was unreachable from this run's sandboxed network (egress policy denial to that host). Verification above was done directly against the current MDX source in the powersync-docs repo instead.
Filed automatically by Claude Code (claude-sonnet-5). A human must verify the scope, prepare the docs PR, and publish it only once this update has been released.
What's changing: PowerSync Service adds a configurable
heartbeat_interval_secondsoption for MongoDB, Postgres, and SQL Server replication sources. For MongoDB sources specifically, the default behavior also changes: if there's no recent resume-token update (typically due to replication lag from bulk writes elsewhere in the cluster), the service now writes a heartbeat to_powersync_checkpointsso the change stream can make progress once it catches up. This is aimed at avoiding thePSYNC_S1345change stream timeout error, which previously could require re-replication from scratch to recover.High priority — existing docs are now inaccurate or incomplete
https://docs.powersync.com/debugging/error-codes) — thePSYNC_S1345entry describes the timeout ("Failed to read MongoDB Change Stream due to a timeout... If this is not resolved after retries, replication may need to be restarted from scratch") but doesn't mention the newheartbeat_interval_secondsoption or the MongoDB default-behavior change specifically designed to reduce/avoid this error.For reviewer consideration — may be worth documenting
https://docs.powersync.com/configuration/powersync-service/self-hosted-instances) — the exampleservice.yamlreplication config skeleton doesn't listheartbeat_interval_secondsas a supportedreplication.connectionsoption. Candidate for addition if the team wants it documented as an advanced/tunable setting; the PR notes the configurable range may still expand over time.Source PR: powersync-ja/powersync-service#747 — merged 2026-08-05.
Note:
docs.powersync.com(llms.txt/llms-full.txt) was unreachable from this run's sandboxed network (egress policy denial to that host). Verification above was done directly against the current MDX source in thepowersync-docsrepo instead.Filed automatically by Claude Code (
claude-sonnet-5). A human must verify the scope, prepare the docs PR, and publish it only once this update has been released.