Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
40 commits
Select commit Hold shift + click to select a range
6c5344f
changes to move to new design
asm582 Jun 26, 2025
820fd30
remove duplicate structs
asm582 Jun 26, 2025
3275a63
add details about inferno
asm582 Jun 26, 2025
42d1add
add actuator interface
asm582 Jun 26, 2025
db7a4d3
update docs
asm582 Jun 26, 2025
f779478
add optimizer folder
asm582 Jun 26, 2025
f8d3d35
remo duplicate API structs
asm582 Jun 26, 2025
ec40337
add modelanalyzer folder
asm582 Jun 26, 2025
afe16e1
add actuator folder
asm582 Jun 26, 2025
5e702d3
updates types.go based on inferno APIs
asm582 Jul 1, 2025
c914a4a
add sample
asm582 Jul 1, 2025
4bd1e63
changes for new API
asm582 Jul 2, 2025
14e96b7
update readme to spawn cluster
asm582 Jul 2, 2025
d21cd32
change install, fix domain and create inventory in reconcile
asm582 Jul 2, 2025
00e568d
wait for deployments to spawn
asm582 Jul 2, 2025
3b02c6f
update deploy/local-cluster.sh to 3-node cluster with nvidia, intel, …
haroldship Jul 2, 2025
b2f89a0
Merge branch 'fake-gpus' into move_ctrl_rntm
haroldship Jul 2, 2025
766c565
add comment for grouping
asm582 Jul 2, 2025
ec49265
always requeue and comments
asm582 Jul 3, 2025
3e57576
use wall-clock time to wake up reconciler
asm582 Jul 3, 2025
6701c45
add configmap to run periodic opt
asm582 Jul 3, 2025
752c173
move collection to collector.go, collect also amd and intel gpu info
haroldship Jul 3, 2025
0035507
add configmap
asm582 Jul 3, 2025
4978600
wait for manager to start
asm582 Jul 3, 2025
3326bbc
enable vllm metrics scraping
asm582 Jul 3, 2025
44155b0
add sample vllm deployment
asm582 Jul 3, 2025
3df22e2
changes to update status
asm582 Jul 7, 2025
496681c
rem duplicate log imports
asm582 Jul 7, 2025
38d11c8
start consuming promethues and query metrics
asm582 Jul 8, 2025
c197cb0
enable vllme
asm582 Jul 9, 2025
6f372fb
remove Collector class
haroldship Jul 9, 2025
4e9aa4e
add cms and API changes
asm582 Jul 10, 2025
af64b47
add dummy optimizer
asm582 Jul 10, 2025
798a047
read slo from cm
asm582 Jul 10, 2025
06d4f98
add actuator
asm582 Jul 10, 2025
e269e26
update readme
asm582 Jul 10, 2025
e1e9193
add owner ref
asm582 Jul 13, 2025
5e4df29
rename api
asm582 Jul 14, 2025
01dbd5f
remove old deployments
asm582 Jul 14, 2025
1d822cc
add collector to its own folder
asm582 Jul 14, 2025
File filter

Filter by extension

Filter by extension


Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
5 changes: 5 additions & 0 deletions .gitignore
Original file line number Diff line number Diff line change
Expand Up @@ -4,6 +4,8 @@
*.dll
*.so
*.dylib
bin/
.vscode/

# Test binary, built with `go test -c`
*.test
Expand All @@ -14,6 +16,9 @@
# Dependency directories (remove the comment below to include it)
vendor/

# Kind config file
kind-config.yaml

# Miscellaneous
.DS_Store
~*
53 changes: 28 additions & 25 deletions Dockerfile
Original file line number Diff line number Diff line change
@@ -1,30 +1,33 @@
# Use a multi-stage build
FROM golang:1.23-bookworm AS builder
# Build the manager binary
FROM docker.io/golang:1.23 AS builder
ARG TARGETOS
ARG TARGETARCH

# Install the lpsolve package
RUN apt-get update && apt-get install -y liblpsolve55-dev
WORKDIR /workspace
# Copy the Go Modules manifests
COPY go.mod go.mod
COPY go.sum go.sum
# cache deps before building and copying source so that we don't need to re-download as much
# and so that source changes don't invalidate our downloaded layer
RUN go mod download

WORKDIR /app
COPY . .
# Copy the go source
COPY cmd/main.go cmd/main.go
COPY api/ api/
COPY internal/ internal/

# Set CGO flags for lpsolve package
ENV CGO_CFLAGS="-I/usr/include/lpsolve"
ENV CGO_LDFLAGS="-llpsolve55 -lm -ldl -lcolamd"
# Build
# the GOARCH has not a default value to allow the binary be built according to the host where the command
# was called. For example, if we call make docker-build in a local env which has the Apple Silicon M1 SO
# the docker BUILDPLATFORM arg will be linux/arm64 when for Apple x86 it will be linux/amd64. Therefore,
# by leaving it empty we can ensure that the container and binary shipped on it will have the same platform.
RUN CGO_ENABLED=0 GOOS=${TARGETOS:-linux} GOARCH=${TARGETARCH} go build -a -o manager cmd/main.go

# Build all main.go files in cmd directory
RUN for file in $(find cmd -name "main.go"); do \
dir=$(dirname "$file"); \
name=$(basename "$dir"); \
go build -o bin/$name $file; \
done
# Use distroless as minimal base image to package the manager binary
# Refer to https://github.com/GoogleContainerTools/distroless for more details
FROM gcr.io/distroless/static:nonroot
WORKDIR /
COPY --from=builder /workspace/manager .
USER 65532:65532

# Create the final image
FROM debian:bookworm-slim
RUN apt-get update && apt-get install -y liblpsolve55-dev
COPY --from=builder /app/bin /bin

# Expose the port the API will listen on
EXPOSE 8080

# Command to run the binary when the container starts
CMD ["optimizer"]
ENTRYPOINT ["/manager"]
225 changes: 225 additions & 0 deletions Makefile
Original file line number Diff line number Diff line change
@@ -0,0 +1,225 @@
# Image URL to use all building/pushing image targets
IMG ?= controller:latest

# Get the currently used golang install path (in GOPATH/bin, unless GOBIN is set)
ifeq (,$(shell go env GOBIN))
GOBIN=$(shell go env GOPATH)/bin
else
GOBIN=$(shell go env GOBIN)
endif

# CONTAINER_TOOL defines the container tool to be used for building images.
# Be aware that the target commands are only tested with Docker which is
# scaffolded by default. However, you might want to replace it to use other
# tools. (i.e. podman)
CONTAINER_TOOL ?= docker

# Setting SHELL to bash allows bash commands to be executed by recipes.
# Options are set to exit when a recipe line exits non-zero or a piped command fails.
SHELL = /usr/bin/env bash -o pipefail
.SHELLFLAGS = -ec

.PHONY: all
all: build

##@ General

# The help target prints out all targets with their descriptions organized
# beneath their categories. The categories are represented by '##@' and the
# target descriptions by '##'. The awk command is responsible for reading the
# entire set of makefiles included in this invocation, looking for lines of the
# file as xyz: ## something, and then pretty-format the target and help. Then,
# if there's a line with ##@ something, that gets pretty-printed as a category.
# More info on the usage of ANSI control characters for terminal formatting:
# https://en.wikipedia.org/wiki/ANSI_escape_code#SGR_parameters
# More info on the awk command:
# http://linuxcommand.org/lc3_adv_awk.php

.PHONY: help
help: ## Display this help.
@awk 'BEGIN {FS = ":.*##"; printf "\nUsage:\n make \033[36m<target>\033[0m\n"} /^[a-zA-Z_0-9-]+:.*?##/ { printf " \033[36m%-15s\033[0m %s\n", $$1, $$2 } /^##@/ { printf "\n\033[1m%s\033[0m\n", substr($$0, 5) } ' $(MAKEFILE_LIST)

##@ Development

.PHONY: manifests
manifests: controller-gen ## Generate WebhookConfiguration, ClusterRole and CustomResourceDefinition objects.
$(CONTROLLER_GEN) rbac:roleName=manager-role crd webhook paths="./..." output:crd:artifacts:config=config/crd/bases

.PHONY: generate
generate: controller-gen ## Generate code containing DeepCopy, DeepCopyInto, and DeepCopyObject method implementations.
$(CONTROLLER_GEN) object:headerFile="hack/boilerplate.go.txt" paths="./..."

.PHONY: fmt
fmt: ## Run go fmt against code.
go fmt ./...

.PHONY: vet
vet: ## Run go vet against code.
go vet ./...

.PHONY: test
test: manifests generate fmt vet setup-envtest ## Run tests.
KUBEBUILDER_ASSETS="$(shell $(ENVTEST) use $(ENVTEST_K8S_VERSION) --bin-dir $(LOCALBIN) -p path)" go test $$(go list ./... | grep -v /e2e) -coverprofile cover.out

# TODO(user): To use a different vendor for e2e tests, modify the setup under 'tests/e2e'.
# The default setup assumes Kind is pre-installed and builds/loads the Manager Docker image locally.
# CertManager is installed by default; skip with:
# - CERT_MANAGER_INSTALL_SKIP=true
.PHONY: test-e2e
test-e2e: manifests generate fmt vet ## Run the e2e tests. Expected an isolated environment using Kind.
@command -v $(KIND) >/dev/null 2>&1 || { \
echo "Kind is not installed. Please install Kind manually."; \
exit 1; \
}
@$(KIND) get clusters | grep -q 'kind' || { \
echo "No Kind cluster is running. Please start a Kind cluster before running the e2e tests."; \
exit 1; \
}
go test ./test/e2e/ -v -ginkgo.v

.PHONY: lint
lint: golangci-lint ## Run golangci-lint linter
$(GOLANGCI_LINT) run

.PHONY: lint-fix
lint-fix: golangci-lint ## Run golangci-lint linter and perform fixes
$(GOLANGCI_LINT) run --fix

.PHONY: lint-config
lint-config: golangci-lint ## Verify golangci-lint linter configuration
$(GOLANGCI_LINT) config verify

##@ Build

.PHONY: build
build: manifests generate fmt vet ## Build manager binary.
go build -o bin/manager cmd/main.go

.PHONY: run
run: manifests generate fmt vet ## Run a controller from your host.
go run ./cmd/main.go

# If you wish to build the manager image targeting other platforms you can use the --platform flag.
# (i.e. docker build --platform linux/arm64). However, you must enable docker buildKit for it.
# More info: https://docs.docker.com/develop/develop-images/build_enhancements/
.PHONY: docker-build
docker-build: ## Build docker image with the manager.
$(CONTAINER_TOOL) build -t ${IMG} .

.PHONY: docker-push
docker-push: ## Push docker image with the manager.
$(CONTAINER_TOOL) push ${IMG}

# PLATFORMS defines the target platforms for the manager image be built to provide support to multiple
# architectures. (i.e. make docker-buildx IMG=myregistry/mypoperator:0.0.1). To use this option you need to:
# - be able to use docker buildx. More info: https://docs.docker.com/build/buildx/
# - have enabled BuildKit. More info: https://docs.docker.com/develop/develop-images/build_enhancements/
# - be able to push the image to your registry (i.e. if you do not set a valid value via IMG=<myregistry/image:<tag>> then the export will fail)
# To adequately provide solutions that are compatible with multiple platforms, you should consider using this option.
PLATFORMS ?= linux/arm64,linux/amd64,linux/s390x,linux/ppc64le
.PHONY: docker-buildx
docker-buildx: ## Build and push docker image for the manager for cross-platform support
# copy existing Dockerfile and insert --platform=${BUILDPLATFORM} into Dockerfile.cross, and preserve the original Dockerfile
sed -e '1 s/\(^FROM\)/FROM --platform=\$$\{BUILDPLATFORM\}/; t' -e ' 1,// s//FROM --platform=\$$\{BUILDPLATFORM\}/' Dockerfile > Dockerfile.cross
- $(CONTAINER_TOOL) buildx create --name inferno-autoscaler-builder
$(CONTAINER_TOOL) buildx use inferno-autoscaler-builder
- $(CONTAINER_TOOL) buildx build --push --platform=$(PLATFORMS) --tag ${IMG} -f Dockerfile.cross .
- $(CONTAINER_TOOL) buildx rm inferno-autoscaler-builder
rm Dockerfile.cross

.PHONY: build-installer
build-installer: manifests generate kustomize ## Generate a consolidated YAML with CRDs and deployment.
mkdir -p dist
cd config/manager && $(KUSTOMIZE) edit set image controller=${IMG}
$(KUSTOMIZE) build config/default > dist/install.yaml

##@ Deployment

ifndef ignore-not-found
ignore-not-found = false
endif

.PHONY: install
install: manifests kustomize ## Install CRDs into the K8s cluster specified in ~/.kube/config.
$(KUSTOMIZE) build config/crd | $(KUBECTL) apply -f -

.PHONY: uninstall
uninstall: manifests kustomize ## Uninstall CRDs from the K8s cluster specified in ~/.kube/config. Call with ignore-not-found=true to ignore resource not found errors during deletion.
$(KUSTOMIZE) build config/crd | $(KUBECTL) delete --ignore-not-found=$(ignore-not-found) -f -

.PHONY: deploy
deploy: manifests kustomize ## Deploy controller to the K8s cluster specified in ~/.kube/config.
cd config/manager && $(KUSTOMIZE) edit set image controller=${IMG}
$(KUSTOMIZE) build config/default | $(KUBECTL) apply -f -

.PHONY: undeploy
undeploy: kustomize ## Undeploy controller from the K8s cluster specified in ~/.kube/config. Call with ignore-not-found=true to ignore resource not found errors during deletion.
$(KUSTOMIZE) build config/default | $(KUBECTL) delete --ignore-not-found=$(ignore-not-found) -f -

##@ Dependencies

## Location to install dependencies to
LOCALBIN ?= $(shell pwd)/bin
$(LOCALBIN):
mkdir -p $(LOCALBIN)

## Tool Binaries
KUBECTL ?= kubectl
KIND ?= kind
KUSTOMIZE ?= $(LOCALBIN)/kustomize
CONTROLLER_GEN ?= $(LOCALBIN)/controller-gen
ENVTEST ?= $(LOCALBIN)/setup-envtest
GOLANGCI_LINT = $(LOCALBIN)/golangci-lint

## Tool Versions
KUSTOMIZE_VERSION ?= v5.6.0
CONTROLLER_TOOLS_VERSION ?= v0.17.2
#ENVTEST_VERSION is the version of controller-runtime release branch to fetch the envtest setup script (i.e. release-0.20)
ENVTEST_VERSION ?= $(shell go list -m -f "{{ .Version }}" sigs.k8s.io/controller-runtime | awk -F'[v.]' '{printf "release-%d.%d", $$2, $$3}')
#ENVTEST_K8S_VERSION is the version of Kubernetes to use for setting up ENVTEST binaries (i.e. 1.31)
ENVTEST_K8S_VERSION ?= $(shell go list -m -f "{{ .Version }}" k8s.io/api | awk -F'[v.]' '{printf "1.%d", $$3}')
GOLANGCI_LINT_VERSION ?= v1.63.4

.PHONY: kustomize
kustomize: $(KUSTOMIZE) ## Download kustomize locally if necessary.
$(KUSTOMIZE): $(LOCALBIN)
$(call go-install-tool,$(KUSTOMIZE),sigs.k8s.io/kustomize/kustomize/v5,$(KUSTOMIZE_VERSION))

.PHONY: controller-gen
controller-gen: $(CONTROLLER_GEN) ## Download controller-gen locally if necessary.
$(CONTROLLER_GEN): $(LOCALBIN)
$(call go-install-tool,$(CONTROLLER_GEN),sigs.k8s.io/controller-tools/cmd/controller-gen,$(CONTROLLER_TOOLS_VERSION))

.PHONY: setup-envtest
setup-envtest: envtest ## Download the binaries required for ENVTEST in the local bin directory.
@echo "Setting up envtest binaries for Kubernetes version $(ENVTEST_K8S_VERSION)..."
@$(ENVTEST) use $(ENVTEST_K8S_VERSION) --bin-dir $(LOCALBIN) -p path || { \
echo "Error: Failed to set up envtest binaries for version $(ENVTEST_K8S_VERSION)."; \
exit 1; \
}

.PHONY: envtest
envtest: $(ENVTEST) ## Download setup-envtest locally if necessary.
$(ENVTEST): $(LOCALBIN)
$(call go-install-tool,$(ENVTEST),sigs.k8s.io/controller-runtime/tools/setup-envtest,$(ENVTEST_VERSION))

.PHONY: golangci-lint
golangci-lint: $(GOLANGCI_LINT) ## Download golangci-lint locally if necessary.
$(GOLANGCI_LINT): $(LOCALBIN)
$(call go-install-tool,$(GOLANGCI_LINT),github.com/golangci/golangci-lint/cmd/golangci-lint,$(GOLANGCI_LINT_VERSION))

# go-install-tool will 'go install' any package with custom target and name of binary, if it doesn't exist
# $1 - target path with name of binary
# $2 - package url which can be installed
# $3 - specific version of package
define go-install-tool
@[ -f "$(1)-$(3)" ] || { \
set -e; \
package=$(2)@$(3) ;\
echo "Downloading $${package}" ;\
rm -f $(1) || true ;\
GOBIN=$(LOCALBIN) go install $${package} ;\
mv $(1) $(1)-$(3) ;\
} ;\
ln -sf $(1)-$(3) $(1)
endef
20 changes: 20 additions & 0 deletions PROJECT
Original file line number Diff line number Diff line change
@@ -0,0 +1,20 @@
# Code generated by tool. DO NOT EDIT.
# This file is used to track the info used to scaffold your project
# and allow the plugins properly work.
# More info: https://book.kubebuilder.io/reference/project-config.html
domain: ai
layout:
- go.kubebuilder.io/v4
projectName: inferno-autoscaler
repo: github.com/llm-d-incubation/inferno-autoscaler
resources:
- api:
crdVersion: v1
namespaced: true
controller: true
domain: ai
group: llmd
kind: VariantAutoscaling
path: github.com/llm-d-incubation/inferno-autoscaler/api/v1alpha1
version: v1alpha1
version: "3"
Loading