v0.8.36

Changes

• run nccl perf test (#617) @philsippl
• chore: Set FAKE_DB_SIZE=6M (#616) @wojciechsromek
• chore: MAX_BATCH_SIZE=64 (#615) @wojciechsromek

v0.8.35

What's Changed

• Update Hawk Pack by @iliailia in #605
• chore: Bump processing timeout by @wojciechsromek in #608
• Batch evaluation of distances in SMPC by @iliailia in #606
• chore: Adjust heartbeat interval in prod by @wojciechsromek in #610
• fix: Staging heartbeat by @wojciechsromek in #609
• chore: Enable NCCL logs by @wojciechsromek in #611
• chore: Scale down prod by @wojciechsromek in #612
• chore: Re-enable and bump max batch size by @wojciechsromek in #613
• chore: Provide heartbeat retries in config by @wojciechsromek in #614

New Contributors

• @iliailia made their first contribution in #605

Full Changelog: v0.8.34...v0.8.35

v0.8.34

Changes

• chore: Bump prod (#603) @wojciechsromek

Bug Fixes

• parse array from config (#604) @philsippl

v0.8.33

What's Changed

• chore: Bump ServerActor timeout by @wojciechsromek in #595
• chore: Bump log level by @wojciechsromek in #596
• chore: Scale down prod by @wojciechsromek in #597
• chore: Scale up prod by @wojciechsromek in #598
• scale pods to 0 by @carlomazzaferro in #600
• chore: Create R53 entries for all participant nodes by @wojciechsromek in #601
• chore: Enable prod by @wojciechsromek in #602
• Update HawkPack version by @rdragos in #593
• simpler heartbeat based on health by @philsippl in #599

Full Changelog: v0.8.32...v0.8.33

v0.8.32

What's Changed

• upgrade prod to latest version by @danielle-tfh in #589
• scale pods to 1 by @carlomazzaferro in #590
• chore: Scale down prod by @wojciechsromek in #592
• expose partial matches by @philsippl in #534
• chore: Run in prod, disable persistence, print info by @wojciechsromek in #594

Full Changelog: v0.8.31...v0.8.32

v0.8.31

Changes

• chore: Scale down prod for upgrade-proto (#588) @wojciechsromek

Bug Fixes

• fix metric being emitted (#587) @danielle-tfh
• use new key for db url (#584) @carlomazzaferro

v0.8.30

Changes

• Add panic logging configuration (#585) @danielle-tfh

v0.8.29

What's Changed

• chore: Try 80 parallel loading threads with Aurora by @wojciechsromek in #581
• Add results and request metrics and update prod by @danielle-tfh in #582
• fix secret key by @carlomazzaferro in #583

Full Changelog: v0.8.28...v0.8.29

v0.8.28

What's Changed

• bump version to v0.8.27 by @eaypek-tfh in #566
• increase parallelism to 16 by @carlomazzaferro in #568
• try 4 parallelism by @carlomazzaferro in #569
• revert parallelism. increase startup probe by @carlomazzaferro in #570
• CPU refactoring by @rdragos in #559
• chore: Use 16 parallel DB loading threads by @wojciechsromek in #571
• chore: Use 4 parallel DB loading threads by @wojciechsromek in #572
• chore: Use 1 parallel DB loading threads by @wojciechsromek in #573
• chore: Use 2 parallel DB loading threads by @wojciechsromek in #574
• chore: Use Aurora DB in participant 1 by @wojciechsromek in #575
• chore: Try 16 parallel loading threads with Aurora by @wojciechsromek in #576
• chore: Try 64 parallel loading threads with Aurora by @wojciechsromek in #577
• increase parallelism to 128 by @carlomazzaferro in #578
• chore: Try 96 parallel loading threads with Aurora by @wojciechsromek in #579
• Add creating global metrics recorder in file by @danielle-tfh in #580

Full Changelog: v0.8.27...v0.8.28

v0.8.27

Changes

• log db load into memory progress (#565) @eaypek-tfh
• bump iris-mpc startup probe (#563) @eaypek-tfh
• Revert "chore: Set replicas to 0 in prod for upgrade (#553)" (#562) @carlomazzaferro
• Add backtrace to be 1 (#561) @danielle-tfh
• enable metrics on staging (#560) @danielle-tfh
• chore: Set replicas to 0 in prod for upgrade (#553) @wojciechsromek
• chore: Disable NCCL logs (#552) @wojciechsromek

Bug Fixes

• ~4x startup probe (#564) @carlomazzaferro