feat(devx): local dev setup for control plane and full end-to-end flow (MLI-6681)

[lilyz-ai]

Summary

Adds a complete local development workflow for model-engine so developers can iterate on both control plane code and the full endpoint lifecycle without cloud credentials or prod images.

Control-plane-only mode (make dev-server):

• Spins up Postgres + Redis via docker-compose
• LOCAL=true activates fake queue/docker/k8s implementations (mirrors CIRCLECI=true)
• Full gateway API available at :5000 with auth skipped — no k8s cluster needed

Full end-to-end mode (make dev-server-full + make dev-service-builder + make dev-k8s-cacher):

• make kind-up + make kind-image creates a local kind cluster and loads model-engine:local into it
• Service Builder picks up endpoint creation tasks from local Redis and creates real k8s Deployments in kind
• K8s Cacher polls kind and writes endpoint status back to Redis
• Echo server (model-engine:local) used as the inference container — no GPU required

Code fixes included:

• service_builder/celery.py + celery_task_queue_gateway.py: onprem cloud provider now uses redis Celery backend instead of s3 — without this, the Service Builder writes results to Redis but the Gateway looks in S3, leaving endpoints stuck in PENDING
• dependencies.py: LOCAL=true + cloud_provider=onprem falls through to real OnPremQueueEndpointResourceDelegate instead of the fake
• env_vars.py: GIT_TAG defaults to "local" when LOCAL=true so k8s templates reference the correct model-engine:local image

New files:

• docker-compose.local.yml — Postgres 15 + Redis 7 with healthchecks and persistent volume
• service_configs/service_config_local.yaml — HMI config for local services
• model_engine_server/core/configs/local-full.yaml — onprem infra config for kind
• Makefile — all dev targets in one place

Test plan

• make dev-up && make dev-migrate && make dev-server — gateway starts, GET /v1/model-endpoints returns 200
• make kind-up && make kind-image — kind cluster created, model-engine:local loaded
• make dev-server-full + make dev-service-builder + make dev-k8s-cacher — all three processes start cleanly
• POST a sync CPU endpoint with the echo server image → pod appears in kubectl --context kind-llm-engine get pods -n model-engine and endpoint transitions to READY
• Existing unit tests pass: make test

Closes MLI-6681

🤖 Generated with Claude Code

Greptile Summary

• Adds a complete local development workflow with two modes: control-plane-only (make dev-server using fake k8s/queue/docker) and full end-to-end via kind (make dev-server-full + service builder + k8s cacher), with all env vars pinned in the Makefile.
• Fixes a backend/broker mismatch for cloud_provider=onprem: both celery_task_queue_gateway.py and service_builder/celery.py now use \"redis\" as the Celery result backend for onprem (previously fell through to \"s3\"), which resolves endpoints getting stuck in PENDING in the full local flow.
• dependencies.py is updated so LOCAL=true + onprem correctly uses the real OnPremQueueEndpointResourceDelegate and Redis task queues, while LOCAL=true + any other provider still uses the fake delegate.

Confidence Score: 5/5

Safe to merge — all previously flagged issues have been addressed and no new P1/P0 issues were found.

All three previously identified issues (cache_redis_aws_url, ML_INFRA_SERVICES_CONFIG_PATH not pinned, celery backend mismatch) are resolved in this revision. The logic changes in dependencies.py, celery_task_queue_gateway.py, and service_builder/celery.py are consistent and correct. Only dev tooling and config files are affected.

No files require special attention.

Important Files Changed

Filename	Overview
model-engine/Makefile	New Makefile with dev targets; ML_INFRA_SERVICES_CONFIG_PATH is now pinned in both LOCAL_ENV and FULL_LOCAL_ENV, and docker compose --wait is used instead of manual polling loops.
model-engine/model_engine_server/api/dependencies.py	Three conditional blocks updated: fake queue delegate is skipped for LOCAL+onprem, Redis task queues selected for LOCAL mode, and FakeDockerRepository used for LOCAL. Logic is correct for both control-plane-only and full e2e modes.
model-engine/model_engine_server/infra/gateways/celery_task_queue_gateway.py	Backend protocol now correctly uses 'redis' for onprem, matching the fix in service_builder/celery.py; resolves the gateway/service-builder backend mismatch that caused endpoints to stick in PENDING.
model-engine/model_engine_server/service_builder/celery.py	Backend protocol updated to 'redis' for onprem; broker type for onprem is driven by celery_broker_type_redis: true in local-full.yaml via force_redis, which correctly yields Redis.
model-engine/service_configs/service_config_local.yaml	Uses cache_redis_onprem_url (not cache_redis_aws_url), correctly bypassing the cloud-provider assertion in hmi_config.cache_redis_url.
model-engine/model_engine_server/core/configs/local-full.yaml	New onprem infra config for kind; celery_broker_type_redis: true is the key flag that routes the service builder broker to Redis instead of SQS for onprem.

_{Reviews (6): Last reviewed commit: "fix(test): include otlp.proto.grpc in OT..." | Re-trigger Greptile}

[lilyz-ai]

@greptile review

[lilyz-ai]

/greptile

[lilyz-ai]

@greptile review

[lilyz-ai]

/greptile