-
Bug
-
Resolution: Done
-
Major
-
7.0.1.GA
-
None
-
-
-
-
-
-
Tested with EAP 7.0.1:
1. Start EAP and connect to CLI
2. /subsystem=undertow/server=default-server/host=default-host/setting=access-log:add(extended=true,pattern="time-taken")
3. /subsystem=undertow/server=default-server/http-listener=default:write-attribute(name=record-request-start-time,value=true)
4. reload
5. deploy attached testing war app long-running-servlet.war
6. perform request on server, e.g.: http://localhost:8080/long-running-servlet/HeavyProcessing?duration=6000 and see content in EAP_HOME/standalone/log/access_log.log fileTested with EAP 7.0.1: 1. Start EAP and connect to CLI 2. /subsystem=undertow/server=default-server/host=default-host/setting=access-log:add(extended=true,pattern="time-taken") 3. /subsystem=undertow/server=default-server/http-listener=default:write-attribute(name=record-request-start-time,value=true) 4. reload 5. deploy attached testing war app long-running-servlet.war 6. perform request on server, e.g.: http://localhost:8080/long-running-servlet/HeavyProcessing?duration=6000 and see content in EAP_HOME/standalone/log/access_log.log file
There is a bug in Undertow's way of converting response time to seconds format for access log. In milis variable is expected to be time in milliseconds but as timeUnit is TimeUnit.SECONS actually, there are just seconds after such conversion.
You can see it for example when use 'time-taken' identifier in access log and perform some long running request - see 'Steps to Reproduce' section.
- clones
-
-
- Resolved
-