Description
There is an integration in which the SOAPProbe response is large (~12mb) and the SOAPProcessorThread associated with processing the payload is not releasing the memory related to the associated business rule. Memory is leaking at a rapid rate and is causing the instance to become slow and require a restart.
There is a scheduled job that daily sends a SOAPProbe message to the output ECC queue. The MID Server picks it up and makes the SOAP request to the 3rd party integration, which returns a 12mb payload (46k+ CI-like rows returned) to the ECC input queue. There are no errors associated with the SOAPProcessorThread, but the ECC queue record is never marked as "Processed" and does not release its memory.
There is a scheduled job that daily sends a SOAPProbe message to the output ECC queue. The MID Server picks it up and makes the SOAP request to the 3rd party integration, which returns a 12mb payload (46k+ CI-like rows returned) to the ECC input queue. There are no errors associated with the SOAPProcessorThread, but the ECC queue record is never marked as "Processed" and does not release its memory.
Steps to Reproduce
- Create a business rule on the ecc_queue table that takes a lot of time to process.
- Send a SOAP request via MID server when the response payload is at least 10MB.
Note SOAPProcessorThread taking memory.
Workaround
Do not create onBefore or onAfter business rules on the ecc_queue table. Use an asynchronous business rule instead.
Related Problem: PRB650037