Open Sourcemerged upstream work, with receipts

PR #1417 Feat/replication audit events: replication tasks now emit aggregated, status-keyed telemetry into a queryable $system bucket, so users can query replication sync state like any other data and build diagnostic dashboards and notifications on top of it, instead of having no visibility. Reviewed and merged by the maintainer, shipping in ReductStore v1.20. Endorsed by the co-founder ↑

Read the PR on GitHub →

PR #1431 Feat/usage statistics: every node now writes its own usage statistics (read/write traffic counted at the storage-engine level plus storage-size snapshots) as queryable $system records on a 60-second interval, with replicas forwarding to the primary for one consolidated, operational view of the whole instance. Reviewed and merged by the maintainer, shipping in ReductStore v1.20.

Read the PR on GitHub →

PR #1474 Feat/per bucket usage stats: usage telemetry gains per-bucket granularity (each bucket emits its own $system event alongside the instance total) plus new written_entries, read_entries, and record_count metrics, so operators can attribute traffic and storage to individual buckets instead of only the instance as a whole. Reviewed and merged by the maintainer, shipping in an upcoming release.

Read the PR on GitHub →

PR #1481 feat(syslog): capture log messages to $system/logs/<instance> (#1467): the database can now capture its own log output as queryable data: an opt-in pipeline writes system log messages into a $system/logs/<instance>/messages stream at a configurable severity level, so a node's logs can be queried as time-series records alongside its other system events instead of only scrolling past on stdout. Reviewed and merged by the maintainer, shipping in an upcoming release.

Read the PR on GitHub →

PR #1496 Feat/unify syslog: all $system event families (audit, usage, replication, lifecycle, and log capture) are now routed through one generic sink in a unified syslog module, with a routing enum resolving each family's $system path and the duplicated writer and path logic collapsed, while characterization tests pin the external record format so nothing user-facing changes. Reviewed and merged by the maintainer, shipping in an upcoming release.

Read the PR on GitHub →

PR #1527 Pipeline replication batch sending (#1061, supersedes #1415): replication throughput improves by pipelining batch sending (while one entry's batch is in flight over HTTP, the next entry's batch is already being prepared) with transaction acks moved into the send task and panic recovery that rebuilds the remote bucket, so an internal bug degrades a single pass instead of crashing the database. Reviewed and merged by the maintainer, shipping in an upcoming release.

Read the PR on GitHub →

PR #1538 Replication: add optional payload compression (#1348): replication batches can now be compressed in transit: a per-task compression setting (zstd or gzip, default none) compresses the batch body using standard HTTP Content-Encoding, with transparent decompression on the receiving router and automatic fallback to uncompressed transfer for destination servers too old to decompress requests. Reviewed and merged by the maintainer, shipping in an upcoming release.

Read the PR on GitHub →

PR #3840 Fix untyped-import false positive for packages with py.typed marker (#3445): Pyrefly flagged a false-positive untyped-import diagnostic for libraries that ship their own types: it recommended installing stub packages even when a PEP 561 py.typed marker was present (e.g. modern requests). The fix detects that marker before warning, walking to the package root so it works for both top-level imports and submodules. Reviewed and merged by the maintainer, shipping in an upcoming release.

Read the PR on GitHub →