I am running a nodejs express backend on fork mode. Concurrent requests get completed initially for half an hour, after which, the backend restarts causing a downtime. Why is it happening?

From oracledb.getPool('MS').getStatistics(); i get to see requestsEnqueued property starts increasing at this time, but never getting dequeued.

This is how my api ends res.json(); (ie. I am not using return; ). Am I not completing the request-response cycle properly? Is it making the request alive in the server ?

You may assume initially concurrent users are able to access the app smoothly. Also, the database pool and ENOUGH connections are being created and managed properly. (Eg. connections are being used and release to the pool)