Skip to main content
nevan
Staff
Staff
March 3, 2025

Technical Tip: 'htx_e_busy' counter detail and reason behind the counter keep increasing in the interface states

  • March 3, 2025
  • 0 replies
  • 1245 views
Description This article describes the counter 'htx_e_busy', the reason the counter appears in the interface states, and the primary workaround for the scenario.
Scope FortiOS.
Solution

The 'htx_e_busy' error is associated with the HTTP/HTTPS proxy or the Web Application Daemon (WAD) process. This counter typically appears when the firewall is overwhelmed with web traffic, indicating that it is struggling to handle requests efficiently.

In general, when the host CPU detects that NPU is busy with packet handling, FortiOS tries to retransmit the packet when the NPU queue is full, the host CPU gets busy and raises the counter. In some cases, the packet can be dropped and cause performance issues in the traffic flow.


The counter can be observed with 'get hardware nic portxor 'diagnose hardware deviceinfo nic portXunder the host counter.


diagnose hardware deviceinfo nic portX

==== Host Counters =============
hrx_pkts :21826682137
hrx_bytes :2133806427427
htx_pkts :1789567396
htx_bytes :323790138988
htx_drop :0
htx_e_busy :83030003847
htx_e_busy :2281333

rx_err :133
rx_crc_err :129


Possible cause behind the counter:

High Traffic Volume – Excessive HTTP/HTTPS requests may overload the NPU system.
Limited System Resources – If CPU or memory valid usage is too high, performance may degrade.
Heavy Proxy Inspection – Features like Deep Packet Inspection (DPI) or SSL/TLS decryption can increase processing demands.
Exceeding Connection Limits – The firewall may have reached its maximum number of allowed concurrent connections.
Firmware Issues – Some versions of FortiOS may have known issues that contribute to this error.

To identify the root cause primarily need to monitor the CPU and memory resources with the 'get system performance status' and to identify resource intensive process 'diagnose sys top'. 

The suggested workaround to reduce the counter could be restarting the WAD process with the CLI command 'diagnose test application wad 99' or rebooting the firewall

If the issue persists after the reboot or restart of the daemon, contact the Technical Assistance Center to check the firmware-related or other issues. 

    Thought Leadership Security Summit. Outpace New Threats with AI - enhanced defense. Tuesday, Septmeber 15, 8:30 AM - 2:30 PM PT. The Golf Club at Newcastle, WA.
    Fortinet Flag the Hack. Wednesday, August 26, 9:00 AM - 5:00 PM ET, COSM, Atlanta, GA.