RUM Waterfall- Request Time always high

[shahrukh_niazi]

In our RUM data , i always see "Request Time" to be always on a very high side , sometime as high as few 3-4sec+ , as compare the "response time" or other metrics

What could be the reasons of high "request times" , & any general recommendations to improve it?

see attach screenshot pls

We are in AWS West & all our users are from US only

[shahrukh_niazi]

One more thing. We do have SSO in the mix which is running MS Azure

[AntonioSousa]

Can you post a screenshot of the waterfall?

[shahrukh_niazi]

Already did in my orignal post

[shahrukh_niazi]

attached

[AntonioSousa · Accepted solution]

@shahrukh_niazi 

Sorry, I didn't get the attachment, as I normally expect an inline image 🙂

From the attachment, and this is what we see in a lot of cases, I believe what is happening is the following:

1. There is a TCP/443 secure connection that is being established, and that took in this case 83 ms. This is the time to be expected for the 1 RTT for the TCP connection and 3 RTT for the secure TLS connection.
2. You've got blocking & application cache times on the client side too
3. There's also callback time from the client side.
4. You've got 678 ms for the request, but on the server side it should have taken about 20 ms less. Doesn't seem you have Purepaths for those requests though.

To improve, I would try to find out why a secure connection is happening in this case? Depending on the connectivity, this is something that even much worse than what is shown in the waterfall.

I would also look at how this is being invoked, as blocking and application cache times are things related to browser behaviour, and how it deals with several types of page construction. Callback time is also related to how these XHRs are being invoked.