Profile Image
PPC

Frequent thread dumps and slowness issues

Hi Team,

Please help analyse the attached thread dumps.

We are frequently facing slowness issues and thread dumps.

Thank you.


Report URL - https://fastthread.io/my-thread-report.jsp?p=c2hhcmVkLzIwMjMvMDMvMjQvVEQ0LnR4dC0tMTYtNDMtNjs7VEQyLnR4dC0tMTYtNDMtNjs7VEQzLnR4dC0tMTYtNDMtNjs7VEQ1LnR4dC0tMTYtNDMtNjs7VEQxLnR4dC0tMTYtNDMtNg==

  • Frequent thread dumps

  • slowness issues

  • Kafka messages

  • top command

  • memory leak

Please Sign In or to post your comment or answer

Profile Image

Boushaba Mohammed

Hello,

 

Its seems that your application publishes (produces) some Kafka messages, it might that the producer takes time and causes the slowness, you can start by setting a timeout for your producers by setting the 'delivery.timeout.ms' property.

The size of the published messages might also cause the Kafka producer to take time...

 

In order to get the full image about what is causing the slowness issues, we need more than threadumps

Can you please follow this steps bellow :

  1. Run the command 'top', this will give you list of process that are running, most probably
    you will see that one process is consuming a lot of CPU, if so run this command 'top -H -p {PID}' where {PID} is the process ID which is consuming the highest CPU.
  2. Based on the output of 'top -H -p {PID}' (you can upload it here so i can take a look), you can identify exaclty which threads are consumming the most CPU.
  3. It's also interessting to see GC log, that helps to see if the slowness is not caused by a memory problem (memory leak), to get GC add this to you arguments :
    •  for Java 8 :  -XX:+PrintGCDetails -XX:+PrintGCDateStamps -Xloggc:<file-path>

    • for >=Java 9:  -Xlog:gc*:file=<file-path>

Once you get your GC log, you can upload it here https://gceasy.io/ and share the report (same as you did with trhreaddump)


 

Profile Image

Ankita

Hi PPC,

 

+ I could see your application is suffering from high cpu-spike. If threads loop infinitely, CPU consumption will start to spike up. Tool is reporting below are the looping threads causing cpu-spike.

 

 - Signal Dispatcher
 - Tribes-MembershipReceiver[Catalina-Channel] - receive0() method in  java.net.PlainDatagramSocketImpl file
 - RMI TCP Accept-6969 - socketAccept() method in java.net.PlainSocketImpl file
 - https-jsse-nio-8681-Acceptor - accept0() method in sun.nio.ch.ServerSocketChannelImpl file
 - epollWait() method in sun.nio.ch.EPollArrayWrapper file - 17 threads are looping
 - tomcat-http--38 - line #162 of com.pegarules.generated.pzMakeRuleRefsForPropertyRef_08010 1_2031393717535455885 file in pzMakeRuleRefsForPropertyRef08_01_01() method

 

+ You can check the CPU Spike section and analyze the stacktrace of threads.

 

 

Got something else on mind? Post Your Question

Not the answer you're looking for? Browse other questions tagged
  • Frequent thread dumps

  • slowness issues

  • Kafka messages

  • top command

  • memory leak