Dependency Version Changes

MIF Helm Charts

Core Components

Infrastructure Dependencies

Bundled as sub-charts in moai-inference-framework:

Highlights

Observability Stack

• Integrated Loki + MinIO + Vector for centralized log collection (#64, #67, #69)
• Added Heimdall TTFT, NTPOT, ITL metric panels to Grafana dashboard (#72)
• Added AMD GPU usage monitor dashboard (#71)

Hardware Support

• Added Node Feature Discovery integration with automatic accelerator detection (#62, #77)
• Added NodeFeatureRule for NVIDIA accelerators (H100 NVL, L40S, and more) (#97, #103)
• Guarded NodeFeatureRule with nfd.enabled condition (#92)

Preset Expansion

• Added presets for DeepSeek-R1 max throughput (#46)
• Added InferenceServiceTemplates for Qwen and other models (#45, #49)
• Added quickstart presets for MI250 / MI300x (#94, #100)
• Added vLLM v0.15.1 and v0.17.0 E2E presets for H100 / H200 (#98, #101)
• Added model and framework fields to preset templates (#84)
• Added PYTHONHASHSEED env var to runtime-base templates (#86)
• Standardized preset naming convention (#75)

Documentation (Website)

• Full Docusaurus documentation site setup (#54)
• Added guides: Getting Started, Features, Operations, PD Disaggregation, Prefix Cache-Aware Routing (#59, #61, #74, #91, #93)
• Added Heimdall plugin reference and multi-model serving guide (#57, #88)
• Added Metrics monitoring and log collection docs (#66, #69)
• Added DeepSeek R1 max throughput blog post (#85)

E2E Testing

• Enhanced E2E testing framework with functional tests (#81)
• Added Heimdall tag support and inference service configurations (#56)
• Refactored E2E workflows for performance and quality benchmarks (#51, #53)

Agent Skills

• Introduced agent skills documentation and guides (#68)
• Added bump-dependency skill (#80)

What's Changed

Full Changelog: v0.3.0...v0.4.0

What's Changed

• MAF-19187: feat(config): add tenstorrent and nvidia devices to nfd rule by @hhk7734 in #33
• MAF-19136: feat(deploy): add NVIDIA GPU to dashboard by @hhk7734 in #34
• NO-ISSUE: feat(deploy): bump odin to v0.6.0 by @hhk7734 in #44

Full Changelog: v0.2.0...v0.3.0

What's Changed

• feat(deploy): add DeepSeek R1 MI300 data-parallel inference preset by @bongwoobak in #12
• MAF-19096: feat(config): add nvida a100 to accelerator rule by @hhk7734 in #13
• MAF-19066: feat(test): add end-to-end testing framework and workflows by @seongsukwon-moreh in #14
• MAF-19066: feat(e2e): enhance end-to-end testing framework and workflows by @seongsukwon-moreh in #15
• MAF-19066: feat(e2e): add pd-disaggregation and run inference-perf afterwards by @seongsukwon-moreh in #16
• MAF-19066: refactor(e2e): update CI workflow by @seongsukwon-moreh in #17
• MAF-19066: refactor(e2e): enhance template rendering and structure for inference services by @seongsukwon-moreh in #18
• NO-ISSUE: chore(workflow): Update pr-title-checker.yaml by @ibpark-moreh in #21
• MAF-19066: fix(e2e): fix github action fail by @seongsukwon-moreh in #19
• MAF-19139: feat(preset): add vllm pd dp bases by @hhk7734 in #22
• MAF-19066: feat(e2e): add S3 configuration for inference-perf by @seongsukwon-moreh in #23
• chore: refactor comment guidelines and add test section by @hhk7734 in #24
• NO-ISSUE: style(e2e): indent go template syntax in yaml files by @hhk7734 in #25
• NO-ISSUE: test(e2e): assume fully controlled cluster by @hhk7734 in #26
• MAF-19141: feat(deploy): conditionally enable ecr-token-refresher and move init to pre-install by @hhk7734 in #27
• NO-ISSUE: test(e2e): streamline gateway setup by @hhk7734 in #28
• MAF-19057: feat(helm): add grafana dashboard by @seongsukwon-moreh in #20
• NO-ISSUE: feat(deploy): bump odin and odin-crd to v0.5.1 by @hhk7734 in #29

New Contributors

• @bongwoobak made their first contribution in #12
• @seongsukwon-moreh made their first contribution in #14
• @ibpark-moreh made their first contribution in #21

Full Changelog: v0.1.0...v0.2.0

What's Changed

• feat(deploy): bump odin and odin-crd to v0.4.0 by @hhk7734 in #7
• chore: add CODEOWNERS by @hhk7734 in #8
• feat(config): add moai-accelerator rule by @hhk7734 in #9
• MAF-19069: feat(deploy): add moai-inference-preset chart by @hhk7734 in #10
• MAF-19069: feat(deploy): bump odin and odin-crd to v0.5.0 by @hhk7734 in #11

Full Changelog: v0.0.1...v0.1.0

What's Changed

• feat(deploy): add moai-inference-framework helm chart by @hhk7734 in #1
• feat(deploy): add odin to mif by @hhk7734 in #2
• feat(deploy): add kube-prometheus-stack to mif by @hhk7734 in #3
• feat(deploy): add keda to mif by @hhk7734 in #4
• chore(workflow): add prod cd action by @hhk7734 in #5
• chore(workflow): fix release action by @hhk7734 in #6

New Contributors

• @hhk7734 made their first contribution in #1

Full Changelog: https://github.com/moreh-dev/mif/commits/v0.0.1