There is some heavy file (in GBs) workers are loading up into the memory, for search, its okay if any worker picks up, but while inserting the new data one worker picking up the request makes the other workers to be in the inconsistent state.
As connectivity to database is not possible, as the file is loading from the disk. Looking for some approach, where I can call all the workers while inserting to reduce the inconsistency.
If gunicorn/uvicorn are not built on these lines, what are the other alternative WSGI servers can be used.