Explanation of waterfall phases

[Yosi_Neuman]

Hi guys

Where can we find an explanation on what are each of the waterfall phases?

For example, what is callback stand for?

We have used this link https://varvy.com/performance/navigation-timing-api.html for explanation but there is no callback.

May be there should be an internal one within Dynatrace documentation ?

Yos

[joel_stuart · Accepted solution]

Hi Yos, there's a table you can refer to under the XHR section of the following documentation page: https://www.dynatrace.com/support/help/user-experi...

To answer your question: Callback time is the time between the start time of first callback and the end time of the last callback (that handled the response of the XHR).

Hope that helps.

Joel

[Yosi_Neuman]

Hi Joel,

Just what I was missing!

Great, thanks a lot.

Yos

[jj_woon]

Hi,

May I know what is meant by blocking under Phases? Couldn't find it in the given documentation link page.

Regards,

JJ

[antonio_villarr]

Hi,

I have another question regarding one of the phases: Blocking. What is related to? I was investigating a user action (load home page) with a duration of ~ 31 s, and 28 s was related to blocking, and 2.2s to processing. How should I interpret this huge blocking time? Or it means I have to open a ticket with support?

Kind regards,

Antonio V.

[gilgi · Accepted solution]

Hi Antonio,

Browsers are limited in therms of the number of parallel connections they can run against a "host" (for example www.cnn.com). By default the number is for most browsers 8 parallel connections. This means that if your page is calling a specific host a lot of times with, for example, a lot of images, the first 8 will start running and the next one will only start once one of the previous is completed.

To this you should add the fact that there are other limitations that affect these connections processing, such as JS execution for example.

When the browser sees the reference to the required image, it adds the request for it into the queue of requests. BLOCKING is the time it waits in the queue until it is actually starting.

Gil.