I have following stack Python3.4+Flask+Celery+Gevent+[Flask-]SQLAlchemy+Psycopg2 and there is a problem with celery worker memory usage. Celery beat execution (it queues tasks every 10 seconds): $ celery beat -A some_module -l info Celery worker execution (every task takes ~5 second): $ celery worker -c 1000 -P gevent -l info -A some_module That's ok! But after 2-3 days uptime celery worker uses 90% of memory. The task is simple: we fetch json data from some REST API service ( requests package) and store it to postgresql DB ( flask_sqlalchemy package). How can I fix this problem?

Python3.4+Flask+Celery+Gevent+[Flask-]SQLAlchemy+Psycopg2 memory leak?

I have following stack Python3.4+Flask+Celery+Gevent+[Flask-]SQLAlchemy+Psycopg2 and there is a problem with celery worker memory usage.

$ celery beat -A some_module -l info

$ celery worker -c 1000 -P gevent -l info -A some_module

That's ok! But after 2-3 days uptime celery worker uses 90% of memory.

The task is simple: we fetch json data from some REST API service (requests package) and store it to postgresql DB (flask_sqlalchemy package).

How can I fix this problem?