Add nodes to haproxy frontend fleet

Log ingestion cannot keep up with log generation rate on the production environment's main haproxy frontend fleet (chef role gprd-base-lb-fe). Every weekday, the haproxy logs fall behind for a few hours.

Spreading the traffic across more frontend nodes would allow td-agent to keep up with log ingestion.

For more details see:

• production#6314 (comment 837943580) - Short-term trend
• production#6314 (comment 837948660) - Long-term trend