v0.53.0

What's Changed

• feat(storage): expose local shard replica placement snapshots by @tdakkota in #874
• perf(recordengine): prune physical integer ranges by @tdakkota in #875
• feat(metrics): expose local read revisions by @tdakkota in #876

Full Changelog: v0.52.0...v0.53.0

v0.52.0

What's Changed

• build(deps): Bump github.com/aws/aws-sdk-go-v2/credentials from 1.20.12 to 1.20.13 in the aws group by @dependabot[bot] in #854
• feat(engine): add opt-in series-aware value compression by @tdakkota in #855
• feat(engine): add opt-in independently readable metric chunks by @tdakkota in #856
• perf(engine): clone chunk decoders only on numeric misses by @tdakkota in #857
• perf(engine): compact decoded block cache indexing by @tdakkota in #858
• perf(block): compact independent chunk directories by @tdakkota in #859
• fix(cluster): require tracked claims for primary writes by @tdakkota in #861
• fix(engine): close rollup buckets before aggregation by @tdakkota in #862
• feat(tenant): configure retention independently per signal by @tdakkota in #863
• feat(retention): expose cutoff and cleanup progress by @tdakkota in #864
• feat(admission): optionally reject expired writes by @tdakkota in #865
• feat(fetch): enforce per-signal retention age on reads by @tdakkota in #866
• fix(router): include retention in write rejection totals by @tdakkota in #868
• perf(recordengine): memoize sidecars by source dictionary id by @tdakkota in #869
• perf(recordengine): reuse resolved streams during part reads by @tdakkota in #870
• perf(recordengine): append accepted batch columns directly by @tdakkota in #871
• perf(recordengine): reuse bounded head buffers and flush metadata by @tdakkota in #872

Full Changelog: v0.51.0...v0.52.0

v0.51.0

What's Changed

• perf(block): reuse cached compressed point frames by @tdakkota in #826
• perf(storage): reserve sidecar encoding buffers by @tdakkota in #827
• perf(storage): bypass build barriers for idle identity pruning by @tdakkota in #829
• perf(engine): shard decoded block cache ownership by @tdakkota in #830
• perf(engine): prefer series-local compaction inputs by @tdakkota in #831
• perf(engine): hand off written granule bounds to queries by @tdakkota in #833
• perf(postings): seek contiguous local ordinals arithmetically by @tdakkota in #834
• fix(engine): skip tenure locks for local maintenance by @tdakkota in #835
• perf(engine): bound streaming count-grid reads by @tdakkota in #836
• perf(engine): reuse detached sample map on flush abort by @tdakkota in #838
• perf(engine): reuse frozen buffers in bounded flush runs by @tdakkota in #837
• fix(storage): coalesce busy background metric flushes by @tdakkota in #839
• perf(engine): adapt sparse series index reads by @tdakkota in #840
• feat(backend): expose metadata-only native size hints by @tdakkota in #842
• fix(engine): account for merge finalization memory by @tdakkota in #843
• fix(engine): budget merge inputs before allocation by @tdakkota in #844
• fix(engine): decline busy independent decode admission by @tdakkota in #845
• ci(test): serialize instrumented coverage packages by @tdakkota in #846
• build(deps): Bump the aws group with 4 updates by @dependabot[bot] in #851
• perf(postings): bracket nearby forward seeks by @tdakkota in #847
• perf(engine): avoid state lock for repair presence reads by @tdakkota in #848
• fix(engine): reuse numeric pins between prefetch and scan by @tdakkota in #849
• perf(chunk): preallocate u128 run encoding by @tdakkota in #852
• fix(file): size range reads from the opened file by @tdakkota in #853

Full Changelog: v0.50.2...v0.51.0

v0.50.2

What's Changed

• feat(recordengine): trace key discovery and prune irrelevant parts by @tdakkota in #814
• perf(engine): admit readers during staged metric writes by @tdakkota in #813
• perf(engine): reuse planned series index ranges during reads by @tdakkota in #815
• perf(engine): stop counts after a complete existence proof by @tdakkota in #816
• perf(engine): globally order metric flush parts for range pruning by @tdakkota in #818
• perf(engine): reduce metric frame amplification for sparse reads by @tdakkota in #819
• perf(engine): reuse immutable decoder metadata within the block-cache budget by @tdakkota in #821
• perf(engine): seek string matcher postings directly by @tdakkota in #822
• perf(engine): add timestamp-only count grids by @tdakkota in #823

Full Changelog: v0.50.1...v0.50.2

v0.50.1

What's Changed

• ci(deps): allow avo generator license metadata by @tdakkota in #803
• perf(storage): coalesce retention rewrites and preserve compaction turns by @tdakkota in #804
• perf(engine): avoid redundant metric flush copies by @tdakkota in #805
• perf(engine): reduce WAL staging lookups and admit readers during encoding by @tdakkota in #806
• perf(pdataconv): hash metrics before materializing point attributes by @tdakkota in #808
• perf(engine): prepare contended metric batches before writer ownership by @tdakkota in #811
• perf(engine): stream sealed heads into bounded parts by @tdakkota in #807
• perf(engine): reuse one decoded identity buffer per WAL batch by @tdakkota in #812

Full Changelog: v0.50.0...v0.50.1

v0.50.0

What's Changed

• perf(fetch): push finite string matcher hints to owners by @tdakkota in #788
• build(deps): Bump the aws group with 4 updates by @dependabot[bot] in #783
• build(deps): Bump go.etcd.io/etcd/server/v3 from 3.7.2 to 3.7.3 by @dependabot[bot] in #786
• build(deps): Bump github.com/maypok86/otter/v2 from 2.3.0 to 2.4.0 by @dependabot[bot] in #784
• test(recordengine): reproduce pool-dependent merge retries by @tdakkota in #790
• build(deps): Bump the opentelemetry-collector group across 1 directory with 2 updates by @dependabot[bot] in #639
• perf(index): compact identities and page series lookups by @tdakkota in #791
• perf(storage): batch tenant-contiguous metric appends by @tdakkota in #792
• perf(engine): bound maintenance and publish outside ingestion locks by @tdakkota in #793
• build(deps): Bump github/codeql-action from 4.39.0 to 4.39.1 by @dependabot[bot] in #782
• build(deps): Bump the opentelemetry-collector group with 2 updates by @dependabot[bot] in #795
• build(deps): Bump github.com/aws/aws-sdk-go-v2/credentials from 1.20.10 to 1.20.11 in the aws group by @dependabot[bot] in #796
• perf(bucketindex): avoid unused ancestry indexes in catalog upkeep by @tdakkota in #797
• perf(storage): reuse streamed frame and attribute buffers by @tdakkota in #798
• perf(recordengine): keep cold neighbors out of compression rewrites by @tdakkota in #799
• perf(compress): reuse bounded encoders and decode scratch by @tdakkota in #800
• feat(otlpdecode): expose native protobuf converters by @tdakkota in #801

Full Changelog: v0.49.4...v0.50.0

v0.49.4

What's Changed

• fix(recordengine): serialize record-key cache fills by @tdakkota in #778
• fix(fetch): preserve absent-label matcher semantics by @tdakkota in #774
• perf(promql): enumerate series metadata without samples by @tdakkota in #775
• perf(cluster): preserve projection and safe fetch limits by @tdakkota in #776
• test(heap): skip allocation guards on macOS and Windows by @tdakkota in #781

Full Changelog: v0.49.3...v0.49.4

v0.49.3

What's Changed

• fix(compaction): merge wide parts in current buckets by @tdakkota in #768
• perf(engine): bound count scratch to timestamp blocks by @tdakkota in #769
• fix(recordengine): account projected reads incrementally by @tdakkota in #770
• fix(recordengine): stop non-reducing compaction retries by @tdakkota in #772

Full Changelog: v0.49.2...v0.49.3

v0.49.2

What's Changed

• fix(recordengine): separate disk targets from merge memory by @tdakkota in #764

Full Changelog: v0.49.1...v0.49.2

v0.49.1

What's Changed

• fix(storage): prevent merge starvation and compress resting records by @tdakkota in #762
• build(deps): Bump the opentelemetry group across 1 directory with 2 updates by @dependabot[bot] in #761
• perf(storage): drain merge backlog between ticks by @tdakkota in #763

Full Changelog: v0.49.0...v0.49.1