[FEA]: Use asynccontextmanager to manage ingest_service life cycle

Is this a new feature, an improvement, or a change to existing functionality?

New Feature

How would you describe the priority of this feature request

Make RedisIngestService lifecycle management more explicit. It can help with redis warm up connections and gracefully shutdown.

Please provide a clear description of problem this feature solves

RedisIngestService.get_instance() creates (lazy) and stores a global object the first time it’s called. Its life cycle management is implicit: creation, initialization, gracefully shutdown.

Proposal: add LifeSpan in with context manager at app level (main.py)

Benefits:

• Creates one RedisIngestService per process and stores it as app.state.ingest_service.
• Attempts graceful shutdown/cleanup on app stop.

Describe the feature, and optionally a solution or implementation and any alternatives

logger = logging.getLogger(name)

.....
@asynccontextmanager
async def lifespan(app: FastAPI):
# Build once per process; app owns the instance
app.state.ingest_service = create_ingest_service()
try:
yield
finally:
svc = getattr(app.state, "ingest_service", None)
if svc is not None:
close = getattr(svc, "close", None)
if callable(close):
try:
maybe = close()
import asyncio
if asyncio.iscoroutine(maybe):
await maybe
except Exception:
logger.exception("Error closing ingest service during shutdown")

.....
app = FastAPI(
title="NV-Ingest Microservice",
description="Service for ingesting heterogenous datatypes",
version="25.6.2",
contact={
"name": "NVIDIA Corporation",
"url": "https://nvidia.com",
},
docs_url="/docs",
)

Additional context

No response

Start in main.py at the FastAPI app construction and trace RedisIngestService.get_instance(), including its initialization and shutdown behavior. Compare that flow with the proposed lifespan context manager and app.state.ingest_service ownership. Done means one service is created per process and cleanup is attempted during app shutdown without breaking startup.