Vue d'ensemble (Overview)

• Global Traffic Manager (GTM) – Distributes traffic to your Automation Suite multi-site deployment. The service must be highly available and immune to any deployment site. Additionally, the GTM must be able to configure the health check to isolate the faulty site quickly. The GTM is not mandatory but is recommended for a quick switchover.
• Équilibreur de charge : chaque site a besoin d'un équilibreur de charge local capable d'équilibrer la charge du trafic vers n'importe quel nœud configuré sur le même site.
• Node – Every site must have at least three servers, whereas agent nodes are optional. Bringing the same set of nodes to each cluster is not mandatory. For example, you must host a group of products requiring 96 (v-)CPU/cores. In this case, your Automation Suite cluster in Site 1 can be formed using three servers and three agent nodes of 16 (v-)CPU/cores each, whereas the Automation Suite cluster in Site 2 can be created using three server nodes of 32 (v-)CPU/cores each. Additionally, your Automation Suite cluster hosted in Site 2 can have fewer hardware resources if you do not configure all the products in multi-site availability.
• Redis – Every site hosts a distributed cache (Redis). An application running in either cluster must read consistent data from the store. In addition, you must configure Redis to run in High Availability mode by providing a two-shard HAA license for each cluster.
• SQL Database – It is recommended to configure High Availability for the SQL Server. The previous image depicts the SQL server configured in Always On Availability Groups with a primary SQL server in Site 1, and at least one secondary physically located in Site 2 with data sync enabled. There is also a SQL listener deployed on top of the SQL server. Both clusters are configured to use the address of the same listener.
• Objectstore – Any files or packages uploaded to products are stored in the objectstore. External objectstore must be configured with an Automation Suite deployment to provide more resiliency to failure.
• Elasticsearch (optional) – Elasticsearch is required to send the robot logs. Elasticsearch’s cross-site replication can be configured using the steps described here or by any other method by Elasticsearch.