Symptoms
Hello Bashar,
For last one week we have been noticing abnormal behavior in both K2 servers in production, every evening 5 pm both servers have high CPU and RAM usage and servers stop responding. When users try to open worklist or any K2 form it either takes 8-10 mins or it just keeps on showing loader icon.
We would like to check if there is any timer job or service running on k2 servers that is causing such behavior. This issue automatically gets resolved after 40 mins to 1 hour and everything comes to normal behavior.
Diagnoses
The respected customer found the reason was a group policy to delete borwser's cache at 5:00
"
Hello Bashar,
It appears the issue was because of a group policy that deleted the browser cache of all users at 5pm. The policy was applied as part of some security requirement but currently disabled. We noticed huge amount of ajax calls being generated from client machine to K2 servers causing this outage.
But now that we have identified the source I would like to understand is the expected behavior, if yes how can we manage our policy without effecting the performance on k2 server. "
my last response was:
"
Since k2 works all the day except at that time when the policy executes The issue here occurs due to the Group policy making it the culprit.
Regarding the cache being cleared The behavior is as expected since ALL the users who need to access worklsits or forms will be re authorized and this will put heavy strain for a while.
Perhaps your respected person can look at something like this with the help of k2 services team:
https://www.brightwork.com/product-blog/warm-up-scripts-resolves-slow-loading-site#.WL-4D1XyiUk
"
Resolution
The respected customer found the reason was a group policy to delete borwser's cache at 5:00
"
Hello Bashar Adas,
It appears the issue was because of a group policy that deleted the browser cache of all users at 5pm. The policy was applied as part of some security requirement but currently disabled. We noticed huge amount of ajax calls being generated from client machine to K2 servers causing this outage.
But now that we have identified the source I would like to understand is the expected behavior, if yes how can we manage our policy without effecting the performance on k2 server. "
my last response was:
"
Since k2 works all the day except at that time when the policy executes The issue here occurs due to the Group policy making it the culprit.
Regarding the cache being cleared The behavior is as expected since ALL the users who need to access worklsits or forms will be re authorized and this will put heavy strain for a while.
Perhaps your respected person can look at something like this with the help of k2 services team:
https://www.brightwork.com/product-blog/warm-up-scripts-resolves-slow-loading-site#.WL-4D1XyiUk
"
