HAProxy Enterprise Documentation 2.2r1


The HAProxy Enterprise Stick Table Aggregator provides real-time tracking of client behavior across a cluster of active-active HAProxy Enterprise load balancers. It does this by aggregating the data of stick tables from each load balancer.

When would you use the aggregator? The table below outlines various common ways to set up your load-balancing cluster and which warrant using aggregation:

Standalone load balancer

If you distribute traffic through a single HAProxy Enterprise node, stick tables help improve security and observability by monitoring client activity in real-time.

Active-standby cluster


In an active-standby cluster, you can define the HAProxy Enterprise nodes as peers. The stick tables of the active node are then propagated to passive nodes, and metrics stay accurate on all nodes.

Active-active cluster without Stick Table Aggregator


In an active-active cluster, based on Amazon Elastic Load Balancer or Route Health Injection, for example, every node would propagate updates to all other peer nodes, by default. This would yield inaccurate metrics.

Active-active cluster with Stick Table Aggregator


To ensure that stick table counters are accurate cluster-wide, you can install HAProxy Enterprise Stick Table Aggregator on a dedicated server. It will collect stick table values from all HAProxy Enterprise nodes, then send an aggregate stick table to all individual HAProxy Enterprise nodes.

You can then retrieve the content of the aggregate stick table from any cluster node, and get compound results you can rely on.

The Stick Table Aggregator can aggregate data from HAProxy Enterprise nodes or other Stick Table Aggregators. You can thus set up single-level or multi-level configurations. This helps manage stick tables from different locations or AWS availability zones, for example.

You can also create a cluster of several Stick Table Aggregators to provide high-availability.

In the next sections, we will guide you through configuring several HAProxy Enterprise nodes and one or several Stick Table Aggregator instances to setup:

  • A single-level configuration

  • A multi-level configuration

  • A highly-available cluster of Stick Table Aggregators

Please follow the sections in order, since each section builds upon the previous one.

See also

Next up