Hybrid cluster architecture for reverse proxies

A non-transitory computer-readable medium storing source code that, when executed by a processor, performs a method of performing auto scaling in a high-availability (HA) cluster, the method comprising:

• receiving, by each member of a plurality of members of a hybrid HA cluster of reverse proxy network security devices, initial configuration information regarding whether the member is to operate in an active mode or in a backup mode, wherein at least two members of the plurality of members are configured to operate in the active mode (“active members”), and wherein at least two members of the plurality of members are configured to operate in the backup mode (“backup members”);
• receiving, by a primary member of the active members, network traffic;
• processing, by the primary member, the network traffic;
• indicating, by each member of the plurality of members, ongoing availability by transmitting heartbeat messages to all other members of the plurality of members, wherein active members other than the primary member include information regarding a traffic statistic relating to network traffic being processed within their heartbeat messages;
• aggregating, by the primary member, traffic statistics extracted from the heartbeat messages over a predetermined period of time;
• when the aggregated traffic statistic is greater than a predetermined auto scale up trigger threshold, causing, by the primary member, a backup member of the backup members to become one of the active members by transmitting a first customized heartbeat message to the backup member directing the backup member to enter the active mode;
• and
• when the aggregated traffic statistic is less than a predetermined auto scale down trigger threshold, causing, by the primary member, an active member of the active members to exit the active members by transmitting a second customized heartbeat message to the active member directing the active member to enter the backup mode.

A hardware system for performing failover in a high-availability (HA) cluster, the system comprising:

• a plurality of members of a hybrid HA cluster of reverse proxy network security devices to operate in an active mode or in a backup mode, wherein at least two members of the plurality of members are configured to operate in the active mode (“active members”), and wherein at least two members of the plurality of members are configured to operate in the backup mode (“backup members”);
• and
• a primary member of the active members to receive and process network traffic,
• wherein the hybrid HA cluster detects existence of a failure scenario of a plurality of failure scenarios involving an active member of the active members, the plurality of failure scenarios including (i) failure of the primary member and (ii) failure of a non-primary member of the active members, and seamlessly fails over from the active member to a backup member of the backup members by causing, by the hybrid HA cluster, the backup member to join the active members by placing it in the active mode,
• wherein the reverse proxy network security devices comprise web application firewalls, wherein the network traffic comprises HyperText Transfer Protocol (HTTP) traffic to or from a web application and wherein said performing, by the primary member, the security screening on at least part of the network traffic comprises filtering, monitoring or blocking the HTTP traffic, and
• wherein the reverse proxy network security devices are each implemented in a form of a blade configured to be installed within a chassis-based system, wherein the active members are associated with a first chassis-based system and wherein the backup members are associated with a second chassis-based system and wherein the plurality of failure scenarios further includes failure of the first chassis-based system.