github GreptimeTeam/greptimedb v1.0.0-rc.2
Release v1.0.0-rc.2
on GitHub

pre-release7 hours ago

Release date: March 11, 2026

This release also includes several important bug fixes for repartition:

  • Reject writes to deallocating regions during region merge
  • Ensure entering the staging phase waits for compaction to complete

Upgrading to this version is recommended.

👍 Highlights

Performance Improvements

  • Dynamic filter pushdown for TopK: Runtime dynamic filters are propagated from DataFusion into Mito region scans to reduce unnecessary IO/CPU; in #7545, ORDER BY end_time DESC LIMIT 10 improved from ~28.86s to ~0.21s on non-indexed timestamps.
  • Pruning improvements: Shared partition pruning and extension-range filter improvements improve scan selectivity by #7635 and #7693.
  • Ingestion and scan path optimization: Prometheus decode and remote-write decode are optimized by #7737 and #7761, and Mito2 Parquet scan is accelerated via min-max caches by #7708.

Key New Features

  • PostgreSQL COPY support for ADBC-postgres path: GreptimeDB supports COPY (query) TO STDOUT WITH (format binary) on the PostgreSQL interface, enabling the ADBC-postgres COPY-optimized query path by #7709.
  • Flow TQL CTE support: CREATE FLOW supports simple TQL CTE patterns, improving flow authoring while preserving predictable execution semantics by #7702.
  • Prometheus 3.x behavior alignment: Selector, lookback, and matrix-selector behavior are aligned for more predictable upgrades and query compatibility by #7671 and #7688.

Notable Bug Fixes

  • Repartition safety during region merge: Reject writes to deallocating regions during region merge by #7694.
  • Staging transition correctness: Ensure entering the staging phase waits for compaction to complete by #7776.

Dashboard Updates

  • Query experience: Added table query sidebar and toolbar/editor UI, with refined pagination.
  • SQL and table-name handling: Improved SQL quoting and parsing for quoted/prefixed table names.
  • Usability and stability: Added database selector support, log export limits, and bigint handling fixes.

🚀 Features

  • feat: reduce unit test suite wall time by @waynexia in #7657
  • feat: support group by op in promql by @waynexia in #7663
  • feat(copy_to_json): add date_format/timestamp_format/time_format for JSON format copy by @linyihai in #7633
  • feat: Implement a shared pruner for partitions in the same scanner by @evenyag in #7635
  • feat: adapt prometheus 3.x matrix selector behavior change by @waynexia in #7671
  • feat: partition rule version validation for writes and staging by @WenyXu in #7628
  • feat: cli tool to repair partition columns mismatch by @MichaelScofield in #7605
  • feat: adapt prometheus 3.x selector and lookback behavior by @waynexia in #7688
  • feat: track flow source tables for TQL and info schema by @waynexia in #7697
  • feat: report flow stats from streaming and batching engines by @waynexia in #7701
  • feat: update dashboard to v0.11.12 by @ZonaHex in #7707
  • feat: add a fallback parameter type inference by reading cast type by @sunng87 in #7712
  • feat: support changing table's append_mode to true by @evenyag in #7669
  • feat: implements anomaly_score_iqr, anomaly_score_mad etc. by @killme2008 in #7681
  • feat: implement postgres copy to stdout by @sunng87 in #7709
  • feat: add a subcommand to bench scan by @evenyag in #7722
  • feat(tracing): propagate mailbox trace context and refine procedure spans by @WenyXu in #7726
  • feat: add extension range api for flat format by @evenyag in #7730
  • feat: pull meta config in frontend during startup by @shuiyisong in #7727
  • feat: improve filter support for scanbench by @evenyag in #7736
  • feat: add COPY support for GCS/Azblob with OSS/GCS/Azblob integration coverage by @WenyXu in #7743
  • feat: adapt new name of holt winters fn by @waynexia in #7700
  • feat: flow tql cte by @discord9 in #7702
  • feat: gc batch delete files by @discord9 in #7733
  • feat: admin gc table/regions by @discord9 in #7619
  • feat: implement last row cache reader for flat format by @evenyag in #7757
  • feat: update dashboard to v0.11.13 by @ZonaHex in #7763
  • feat: add function rewrite rule for json_get with cast by @sunng87 in #7631
  • feat: fast path for empty selection files by @waynexia in #7780
  • feat: flat read path support primary_key format memtables by @evenyag in #7759
  • feat: replace shadow-rs with self-maintained version info by @waynexia in #7782
  • feat: use dyn filter by @discord9 in #7545

🐛 Bug Fixes

  • fix: incorrect column statistics of promql plans by @waynexia in #7662
  • fix: drop rhs columns on promql filter join by @waynexia in #7665
  • fix(meta-srv): force close postgres client on reset to release advisory l… by @YZL0v3ZZ in #7673
  • fix: cargo nextest run -p query --features enterprise by @fengys1996 in #7683
  • fix: bump prometheus to 0.14 by @evenyag in #7686
  • fix: panic when querying slowlog by @MichaelScofield in #7689
  • fix(mito2): filter extension ranges in pruner by @fengys1996 in #7693
  • fix(mito2): introduce PartitionExprChange in staging flow and keep memtables on metadata-only updates by @WenyXu in #7695
  • fix: incorrect-tql-explain result by @fengys1996 in #7675
  • fix: use full DDL of flow in information_schema.flows.flow_definition by @sunng87 in #7704
  • fix: handle scalar result in MultiDimPartitionRule by @v0y4g3r in #7715
  • fix(repartition): reject writes on deallocating regions during region merge by @WenyXu in #7694
  • fix: gc schd parallel send hb instr by @discord9 in #7716
  • fix: Install uv to avoid failure while creating databases by @v0y4g3r in #7741
  • fix: preserve the order of primary key indices as defined in the original table info by @WenyXu in #7740
  • fix: null first for part expr as logical expr by @discord9 in #7747
  • fix: Allow overriding sequence when writing flat SSTs by @evenyag in #7764
  • fix(mito2): avoid parquet redownload when write cache already contains file by @WenyXu in #7777
  • fix(meta): remap route addresses when reading full table info by @WenyXu in #7778
  • fix(mito2): ensure enter staging waits for compaction by @WenyXu in #7776

🚜 Refactor

  • refactor(vector-index): use protobuf for metadata and align code by @killme2008 in #7648
  • refactor: remove the RawTableMeta and RawTableInfo to make codes more concise by @MichaelScofield in #7626
  • refactor: rename partition rule version to partition expr version by @WenyXu in #7696
  • refactor: remove session from common meta by @sunng87 in #7698
  • refactor: remote write related code by @v0y4g3r in #7775

📚 Documentation

⚡ Performance

🧪 Testing

  • test: adds sqlness test for vector index by @killme2008 in #7634
  • test: add repartition fuzz target with partition IR and translators by @WenyXu in #7666
  • test(fuzz): repartition validation and add dedicated CI GC profile by @WenyXu in #7703
  • test: fix unstable index meta list test by @evenyag in #7774

⚙️ Miscellaneous Tasks

New Contributors

All Contributors

We would like to thank the following contributors from the GreptimeDB community:

@MichaelScofield, @WenyXu, @YZL0v3ZZ, @ZonaHex, @discord9, @evenyag, @fengys1996, @killme2008, @linyihai, @shuiyisong, @sunng87, @v0y4g3r, @waynexia

Check out latest releases or
releases around GreptimeTeam/greptimedb v1.0.0-rc.2

Don't miss a new greptimedb release

NewReleases is sending notifications on new releases.

Get notifications