SSH Proxy decryption disables vulnerability protection?

[grenzi]

Hello everyone,

  I'm doing some tests with decryption and vulnerability protection. I configured NAT and security policies to permit ssh access to an internal ssh server from the outside and I attached a vulnerability protection profile to the policy. In the vulnerability protection profile I set an exception to block bruteforce login attacks after 2 failed attempts:

This configuration works fine and connections are reset after 2 failed login. Anyway, if I add a SSH decryption policy (SSH proxy) to this connection, the bruteforce login signature is no more triggered by the failed login attempts, so it is no more effective.

Is this an issue or is it "by design", due to the fact that the firewall acts as a man in the middle for the ssh connections?

[mivaldi]

That's an interesting find. Please open a case with Support.

[grenzi]

I opened a case with the support and they confirmed this issue. Waiting for their research and feedback.

Thanks.

[grenzi]

The engineering team confirmed that this is an expected behavior. This is their explanation:

In case SSH-proxy is enabled, the signature 31914 will act on decrypted traffic. Hence, this sig will not trigger because it can not detect the pattern of the failed authentication in the decrypted traffic.

---

[GM001]

I've got an interesting variation on this issue. I have an SSH proxy that inspects all incoming SSH traffic going to various SFTP servers. I have one particular SFTP server that has a security rule that allows SSH, but this one rule ( to just one server) also has a VPP with the same 40015 exception with the action set to alert. Despite the SSH proxy rule covering all incoming SSH, any SSH traffic to that server is showing as NOT decrypted. All the other traffic hitting any other security rules without that exception shows as decrypted. Would adding the Brute Force exception cause the firewall to not decrypt the traffic so the exception can work?

[TomYoung]

Hi @GM001 ,

SSH proxy only allows the detection and block (if configured in the security policy) of the ssh-tunnel App-ID.  https://docs.paloaltonetworks.com/pan-os/9-1/pan-os-admin/decryption/decryption-concepts/ssh-proxy  "The firewall can’t decrypt traffic inside an SSH tunnel."  It only identifies the channel type.

Thanks,

Tom

[GM001]

Thanks Tom, but I'm not sure how that's relevant to my issue. I'm trying to understand how a VPP exception in a security rule can stop an SSH decryption rule from applying.