Certificate for Sophos Web Protection Warnings?

Hello Sophos ma-edv,

Thank you for reaching out to the Sophos Community. Sorry to inform you that with Sophos Endpoint's Web Control components as they are right now, it is not possible to display the "Website Blocked" page for HTTPS sites.

With HTTP traffic this is possible as the connection is not secured. Sophos' Web Control components can inject the "Blocked" webpage into the HTTP traffic so that the message is displayed to inform the end-user that the page was blocked. For HTTPS traffic, as the traffic is secured, this is not possible to do. 

The UTM or XG are able to perform this operation differently. The network appliances do not need to inject the webpage into the traffic, the appliances will instead perform a redirect to their own secured webpage instead of returning the original traffic that was destined for the endpoint. You can picture this as the endpoint connecting to the UTM/XG's webpage displaying a "Block" message. 

If you would like to see changes made to Sophos AV going forward so that Web Control behaves more similarly to the XG or UTM I recommend submitting your vote on the following feature request.
- https://ideas.sophos.com/forums/285723-endpoint-protection/suggestions/18971314-block-page-notice-for-https-content

The following article does a great job of explaining the mechanisms at play in terms of how the XG handles this. You'll find that there is much that needs to be present on the network appliance in terms of how it handles these connections in order to present the block message.  
- https://support.sophos.com/support/s/article/KB-000038420?language=en_US

Thank you very much for the response. I am a still surprised to understand that such an obvious thing seems to require a user-based-feature request. To be honest -especially after your answer - I am not fully convinced that SOPHOS product management has the same level and quality of understanding than the technical support & field engineers do.

You're absolutly right about the complex issues arrising from the challenge to display some kind of error message WITHIN the browser. Here you have all the issues with HTTPS traffic, CA and such.

Anyhow - as we are speaking about Sophos Central - and a windows only environment - THERE IS an Sophos application icon running in the tray. And as this application is intercepting the traffic, it would be for sure capable of displaying some kind of information to the user.

Maybe I am missing something, as I see popup messages in the screen o this KB entry.
https://support.sophos.com/support/s/article/KB-000035134?language=en_US

But I have no clue, how to enable those popup messages within Sophos Central Endpoint Protection.. Any hints?

Hello ma-edv,

this article is for the (near EOL) on-premise SESC. This behaviour has been removed from the Central version, the explanation is in the last paragraph of Rich Baldry's comment.

Christian

The new version of Web Protection/Control is shortly to be available in EAP which does decryption of HTTPS traffic.  As a result this will be able to inject messaging into the browser and behave like HTTP traffic does today.

Important Changes to the Endpoint/Server Protection and EDR Features Early Access Program - Announcements - Endpoint EAP - Sophos Community 

Where should one be able to inject  a message into the browser? Sounds like a marketing myth... Also the link you mentioned doesn't provide any information about the feature to inject a message or even HOW to achieve it.

I am suggesting that once the new web protection/control version is released it will be able to inspect HTTPS traffic. As a result it will be able to display block/warn messages as the current version does today.

Hello  ma-edv ,

not a myth (as others, AFAIK, already do it).  Sorry if stating the already obvious. Traffic redirection is there, and interception, scan and/or injection not a challenge when HTTP is used.  While you can read HTTPS traffic you can't modify it as you'd need the sending server's private key. The solution is to "cut" the connection in two, On the one from/to the browser the server must be mimicked, and for this a root CA must to be installed that is used to issue the pretended server certificate.

Christian    .

Fully agree. No one denys that from a technical prospective; we know this "old school feature" from the UTM anyways. But User 930 was stating that there will be that functionality in the new Web Protection/Control.

But at this time, there is no official annoucment from Sophos about introducing it as a "brand new" feature

By the way, this is not entire true anymore. While old school proxies do this (split into two sessions), DPI engines like SFOS (XGS) are actually using the stream. Therefore the packet will be manipulated while flowing through the firewall and not building up a own connection. But the premise is still the same, you have to trust the certificate anyway from the firewall.