Lately I noticed that free disk space on my 12c Cloud Control server was reducing for some amount. When I investigated, I've figured it out that there were daily incidents of BEA-337 errors with around 40 MB of log  files in /u01/app/Oracle/Middleware/oms12c/gc_inst/user_projects/domains/GCDomain/servers/EMGC_OMS1/adr/diag /ofm/GCDomain/EMGC_OMS1/incident directory.

Problem Key: BEA-337 [WebLogicServer]
Error Message Id: BEA-337

Incident detected using watch rule "StuckThread":
Watch ServerName:       EMGC_OMS1
Watch RuleType:         Log
Watch Rule:             (SEVERITY = 'Error') AND ((MSGID = 'WL-000337') OR (MSGID = 'BEA-000337'))
Watch DomainName:       GCDomain
Watch Data:
   MESSAGE : [STUCK] ExecuteThread: '7' for queue: 'weblogic.kernel.Default (self-tuning)' has been busy for "611" seconds working on the request "weblogic.servlet.internal.ServletRequestImpl@70733a9e[
POST /em/websvcs/emws/ConsoleJobStepExecutorService HTTP/1.1

As you can see, incidents was caused due to a stuck thread watch rule. It was triggered because of "Refresh From My Oracle Support" job which is running more than 600 seconds. First thing I tried was rescheduling the job. From OEM, I navigated to jobs and found the job through advanced search by selecting ”Refresh From My Oracle Support” from the job type drop down and rescheduling it to some less busy time.

However it didn't work out the problem. So I decided to increase the parameter controlling stuck thread watch rule which is StuckThreadMaxTime. It's default value is 600 (seconds) and my job takes 13.5 minutes to complete. 

To alter the parameter:

1. Log on to WebLogic administration console through EMGC_OMS1 target on OMS or directly navigating to https://<host>:7200/console 

2. Click on EMGC_OMS1 under GCDomain->Environment -> Servers. Find "Stuck Thread Max Time" setting on Configuration/Tuning tab. I've increased it to 1200 seconds to cover job duration.

3. Save the configuration and then restart EMGC_OMS1 server from WebLogic administration console or from server console.

This workaround did the job.