topic Re: DEM Server Time vs PurePath Response/Processing in Real User Monitoring

DEM Server Time vs PurePath Response/Processing

serhat_balik — Tue, 02 Feb 2021 06:48:38 GMT

There is a discrepancy between RUM Server Time and PurePath data. You can see that it shows 17 seconds Server Time in waterfall analysis, but when we dive into PurePath of this action, we only get 3.53 ms.

Is there something missing in the middle or am I trying to analyze this in a wrong way?

Re: DEM Server Time vs PurePath Response/Processing

skrystosik — Tue, 02 Feb 2021 09:21:41 GMT

What do you see in waterfall? You can have not all purepath collected for this user action, ot problem was somewhere between request and response and it wasn't caused by server performance but browser or network.

Sebastian

Re: DEM Server Time vs PurePath Response/Processing

Julius_Loman — Tue, 02 Feb 2021 11:20:20 GMT

Apart from the waterfall Sebastian mentions, is there anything between the application server and client apart from the network? (Load balancer / reverse proxy / etc).? For example reverse proxy running on low connection pool.

I remember sometimes browser reporting server time instead of network. The categories (network/server/client) is reported by the user's browser and may not be precise.

Re: DEM Server Time vs PurePath Response/Processing

serhat_balik — Tue, 02 Feb 2021 11:50:45 GMT

We're trying to get and analyze TCP dumps and check where this time is spent. Do you guys have any suggestions about the way we do it?

Re: DEM Server Time vs PurePath Response/Processing

skrystosik — Tue, 02 Feb 2021 13:08:16 GMT

If you are able to reproduce it you can always use Wireshark or similar tool for collection of data.

Sebastian