Release v0.6.0 - Introducing optimal policy to balance read latency and write throughput · zuston/riffle

Highlight

Using the direct write + buffer read to be as optimal policy to balance read latency and write throughput
Introducing the release CI workflow to publish actifacts that has supported the linux amd64 + arm64

feat(ci): Enable memory-prof feature when releasing by @zuston in #65
fix(grpc): Correct status code of huge partition requirement rejection by @zuston in #66
feat(ci): Build binary that run on anolis aarch64 by @zuston in #67
feat(direct-io): Introduce config option to control read/append direct io whether to enable by @zuston in #72
fix(single-buffer-flush): Avoid using the global lock to skip some single partition flush logics by @zuston in #75
fix: Unset runtime blocking thread num to activate parallel ability by @zuston in #76
feat: Trigger binary publish on master push by @zuston in #78
feat: Inject cpu_arch into metric system and http service by @zuston in #79
fix: Add local_disk_healthy metric by @zuston in #81
feat: direct_io with timeout mechanism to fast fail when io hang by @zuston in #80
fix: Fast fail on throwing dir_not_found to ignore events dropped metrics by @zuston in #84
fix: Respect read/append independent option by @zuston in #87
fix: Test failure of get_data_from_localfile by @zuston in #90
fix: Enable O_DIRECT + O_NOATIME flags on direct read by @zuston in #88
debug: Log duration when getting local data + index by @zuston in #91
[followup] debug: Correct partitionId to log by @zuston in #92
feat: Sync all on buffer write to avoid latter page cache flush back storm by @zuston in #93
fix(io-scheduler): Exceed capacity permits request to fallback to the biggest capacity buffer by @zuston in #95
feat: Add the metric of resident bytes by @zuston in #97

Full Changelog: v0.5.0...v0.6.0