We have set up Nginx (on version 1.12.2) webserver. Due to some reason the CPU utilization on the web server is going upto 100% with just 1 or 2 test users in few minutes. This is causing the server to hang.
Could it be caused by GC collection ?
Solved! Go to Solution.
I wouldn't say CPU utilization indicates garbage collection issues, but it still could be something to take a look at.
If garbage collection is your concern you can go to Featured Dashlets > Transaction Flow and filter for the host you expect is experiencing this issue, along with the time frame.
If you hover over the host, you will see a window pop up, GC should be on there and it should be red if there's an issue. If you click on Show Application Process, you should see what garbage collection looks like, along with some other key metrics.
Let me know if you have any questions.
The garbage collection is not an issue here.
So when we try running the application on the nginx webserver without dynatrace agent, the application works fine. The dynatrace webserver agent currently installed is '126.96.36.1999'. We tried installing the latest version 188.8.131.529 but the collector is running a “better”
version therefore bootstrap downloaded the newest version and adjusted the settings for it with version 184.108.40.2069. However we still have the issue of NGINX going upto 100%
CPU and the only way is to restart the service.
Any thoughts on this issue would be helpful.
If the high CPU utilization only occurs with the agent online, you may want to submit a support ticket.
Alternatively, you could try opening up the Purepaths dashlet for a certain time frame and the desired agent, then adding CPU sum to the table and sorting by highest first.
It is possible that a transaction(s) is allocating CPU and never terminates which could lead to CPU starvation. Any Purepaths with high CPU sum would be worth looking into.
Hope that helps.