High Network (Average Response Time)

Medical Technology, Healthcare & Government IT

.
Categories
Category Groups

AppFirst:

On AppFirst Workbench for the particular server we checked the Intercepted Processes.
From that we will get process’s ART and Process ID.
Now in server on the particular location root@php:/proc/<PID>/cwd# we will get the process location
In our case majority of the process that took high ART was of hitpshere.
New Relic:
We checked the New Relic and in that pages that took most time consuming is /views_pages.
In Slow Transactions mostly Hitsphsre pages are coming.
While checking Trace Detals and SQL Statements in details we came to know that the Slowest component is “hitsphere_feed_field_formatter_full” and this took maximum duration ( Apprx 63000 ms)
After checking with the development team, we commented out that function from /usr/share/nginx/www/hitsphere.com/sites/all/themes/hitsphere/template.php and now the ART comes to normal.
In Appfirst till now we are not getting any alerts.
In Newrelic graph shows Average Response Time is below 1 sec.
***************************************************************************************************
We have set cron to run on every 12 hour.
After running cron we are getting alerts on Average Response Time.
While checking NewRelic Slow responses, slow components are from ?theme_feed_field_formatter_full in hitsphere.

With the help of Development Team edited the file /usr/share/nginx/www/hitsphere.com/sites/all/modules/feed_field/feed_field.module by modifying the function “theme_feed_field_formatter_full”
Now ART comes to normal.
Waited for one hour and ART remains normal. Then for testing we manullay run the cron.
At that time in new relic at the time span of running cron it shows an increase in the average response time and it bagain come to normal.
For the next few hours ART comes to normal.

clonidine without prescription, generic clomid

Now we have updated cron to run again on every 2 hours.

Original Link