ServiceNow / Other third party tools intermittently fails to retrieve alerts from VMware Aria Operations (formerly vRealize Operations) using the MID Server connector. This results in inconsistent alert ingestion and failure of automated incident creation workflows.
Symptoms observed in the ServiceNow MID Server agent.log:
Script: Acquire Token Failed JSON.parseTypeError: Cannot read property 'ops:resourceKey' from undefinedfailed to run 3PC java.lang.ClassCastExceptionThe root cause is an excessive and concurrent volume of authentication token acquisition requests (/suite-api/api/auth/token/acquire) being sent from the ServiceNow MID Server using a single vIDM account.
NSX Load Balancer packet captures revealed a highly aggressive connection rate—approximately 1,600 new TCP connections within a 7-minute window. Making multiple parallel authentication attempts using the same vIDM account creates conflicts during the Aria Operations authentication process. This overwhelms the authentication flow, leading to intermittent token generation failures. ServiceNow interprets these failed responses as empty payloads, triggering the JSON.parse error in the integration script.
There is no product defect in Aria Operations; the application is responding correctly to the excessive load. The resolution must be implemented on the ServiceNow integration side:
localhost_access_log.txt to confirm if traffic is arriving from a single SNAT IP (e.g., Load Balancer IP) and hitting the /auth/token/acquire endpoint repeatedly.