A large enterprise has deployed two Palo Alto Networks PA-5250 firewalls in active/passive HA mode with Panorama for centralized management. The network contains over 10,000 users across multiple sites. Recently, the security team deployed a new security policy rule to block a set of high-risk applications. After the commit, the firewall's CPU utilization spiked to 95% and sessions started to drop intermittently. The firewall logs show a high number of session setup failures and timeouts. The existing security policy contains over 5,000 rules. The new rule uses application-based filtering and is placed near the top of the rulebase. What is the most effective course of action to reduce CPU load while maintaining security?
Threat Prevention profiles are more efficient for blocking known applications and offload processing from the policy engine.
Why this answer
The CPU spike is caused by the firewall having to perform application identification (App-ID) on every packet matching the new rule, which is placed near the top of a 5,000-rule policy. By moving the blocking logic to a Threat Prevention profile, the firewall can block the applications at the security profile stage after a faster initial match, reducing the per-session processing overhead. This approach offloads the heavy App-ID processing from the rulebase to a profile that is applied only to relevant traffic, thus lowering CPU utilization while still blocking the high-risk applications.
Exam trap
The trap here is that candidates often assume moving a rule to the bottom (Option B) reduces CPU load by reducing matching frequency, but in reality, App-ID processing is triggered per session regardless of rule position, and the firewall must still evaluate all rules above it, so the CPU spike persists.
How to eliminate wrong answers
Option B is wrong because moving the rule to the bottom of the rulebase does not reduce the CPU load from App-ID processing; the firewall still must perform application identification for every session that matches earlier rules, and the new rule would still require App-ID when reached, potentially increasing latency for matching traffic. Option C is wrong because converting to a service-based filter would bypass App-ID entirely, but it would also fail to accurately block the specific high-risk applications, as many modern applications use non-standard ports or dynamic port ranges, weakening security. Option D is wrong because increasing the session table size and adjusting TCP timewait timeouts addresses symptoms of session setup failures but does not reduce the CPU load caused by App-ID processing; it may even worsen the problem by allowing more concurrent sessions to be processed with the same high overhead.