Change log

Features

• https://github.com/qdrant/qdrant/pull/4254 - Add pyroscope integration for continuous profiling on demand

Improvements

• https://github.com/qdrant/qdrant/pull/4309 - Allow to configure default number of shards per node
• https://github.com/qdrant/qdrant/pull/4317 - Allow to overwrite optimizer settings via config
• https://github.com/qdrant/qdrant/pull/4312, https://github.com/qdrant/qdrant/pull/4369 - Improve vector size estimations, making index thresholds more reliable
• https://github.com/qdrant/qdrant/pull/4428 - Improve default maximum segment size, base it on number of CPUs used for indexing
• https://github.com/qdrant/qdrant/pull/4370 - Use consistent RocksDB settings for both put and remove
• https://github.com/qdrant/qdrant/pull/4376 - Improve ordering of insertions and deletions in RocksDB
• https://github.com/qdrant/qdrant/pull/4371 - Log error if segment flushing failed on drop
• https://github.com/qdrant/qdrant/pull/4352 - Promote REST request processing problems from warning to error
• https://github.com/qdrant/qdrant/pull/4368 - Improve error messages in cases of missing vectors
• https://github.com/qdrant/qdrant/pull/4391 - Improve shard state log message, not strictly related to snapshot recovery
• https://github.com/qdrant/qdrant/pull/4414 - Improve Dockerfile, don't invalidate caches each commit and allow debug settings

Bug fixes

• https://github.com/qdrant/qdrant/pull/4402 - Fix deadlock caused by concurrent snapshot and optimization
• https://github.com/qdrant/qdrant/pull/4411 - Fix potentially losing vectors on crash by enabling RocksDB WAL
• https://github.com/qdrant/qdrant/pull/4416, https://github.com/qdrant/qdrant/pull/4440 - Respect max_segment_size on data ingestion with optimizers disabled, create segments as needed
• https://github.com/qdrant/qdrant/pull/4442 - Fix potentially having bad HNSW links on multithreaded systems