Question about RUM option "Cache control header optimizations"

[ntanaka]

Hi team,

I have a question about this option;

Application settings - Advanced Setup - Cache control header optimizations

It says "If you disable this setting, these headers won't be modified and any changes to application settings that affect your real user monitoring JavaScript code may take a long time to go into effect"

- Question

I wonder how long time it takes to update the JavaScript according to the new RUM configurations.
I am not familiar with cache-control, but I think it will be updated when we delete the browser cache.
Is it correct?

- Background

Our customer is suffering session time out problem when enabling Dynatrace RUM.
We discuss this problem on the support ticket SUP-19637.
We doubt that session time out is caused by cache-control invalidation because it happens when the response header contains the following value;

Name	Value
cache-control	no-cache, no-store
expires	Thu, 01 Jan 1970 00:00:00 GMT

So, we would like to try to disable Dynatrace cache control header optimizations.

Best Regards,
Natsumi Tanaka

[Julius_Loman · Accepted solution]

No problem if you disable the "Cache control header optimizations". If you enable it (by default it is enabled) Dynatrace adds an entry to the Etag HTTP header and will update this Etag with each modification of the application configuration. So if you make any modifications to the Application configuration (such as user tag definition), it may take some time to propagate these changed to clients.

Actually, we have encountered an issue with the Cache control since application returned the Etag header but it was empty (just a single space). Since Dynatrace added another ID to the Etag header, the returned response had Etag set and this led to serious issues with application responses being cached. Unfortunately, this behaviour was caused by a badly designed application (It should not return Etag header with an empty value).

[ntanaka]

Hi Julius,

Thank you very much for your answer.

Now I notice Etag that seems to be added by Dynatrace on the Fiddler trace captured when session time out happened.

Since the etag is not empty, it may not be a problem you mentioned.

Anyway, your information is so helpful for me to understand what Dynatrace add on the response header.

Best Regards,

Natsumi Tanaka