I had a worker service constantly running into an OOM (out of memory) as soon as large amounts of data were processed. the service simply stopped/crashed and did not respond. after outsourcing the large datastructures into temporary postgresql tables it never crashed again.