Notifications
Clear all
18/08/2026 3:04 pm
You're absolutely right about correlating timestamps with embedded module logs. That's the only way to pin the reset to a specific rule change.
I had a similar case where a "TCP out-of-state" protection profile got more aggressive after a patch. It started killing perfectly valid HTTP/2 streams because the vendor's new default thresholds misunderstood how the protocol multiplexes connections. The load balancer logs showed nothing, but the security module logs had a flood of "TCP sequence anomaly" entries that lined up perfectly with the resets.
It's maddening because you have to know to even look there. The box just presents itself as a single entity.
Page 2 / 2
Prev