Skip to Content

Failover and Scalability of SAP Cloud Platform Integration

Hello Colleagues !

In this blog, we shall look into how an SAP Cloud Platform Integration (f.k.a HCI) landscape meets the expectations of a cloud-based application. We shall look into the failover and scalability component of SAP Cloud Platform Integration. This blog is part of the architectural overview of SAP Cloud Platform Integration. You can find all the blogs over here.

I refer to the node architecture of SAP Cloud Platform Integration in this blog. So, it you are new to the concept, I strongly recommend you read this blog before reading further.

Fail-over Capability of SAP Cloud Platform Integration

If the primary instance of a worker node fails to process messages, the failover capability kicks in. Unexpected single instance losses can be due to hardware, system or load failure. SAP Cloud Platform Integration has been designed to be robust against these failures.

SAP Cloud Platform Integration can detect failures, gracefully handle failures, bring up new instances and redirect traffic from a downed instance to a live instance. The tenant management node supervises the worker nodes and shall immediately know if a worker node is not functioning. Health-check reports of worker nodes causes the tenant management nodes to take action.


A pictorial representation can help understand the failover capability of SAP Cloud Platform Integration.




Scalability of SAP Cloud Platform Integration

Assume that you are running an Integration worker node and you are starting to get a high enough level of traffic and load. Additional integration worker nodes are then configure to handle the load. The BigIP load balancer shall spread the load between all the instances. In this situation, you would serve the same integration artifacts from each of the instances – this is ensured by the tenant management node.

In case you face issues achieving the targeted throughput or you observe messages failing during peak loads, please raise a request to LOD-HCI-PI-OPS. We will check the resource consumption of the CPI tenant and adapt accordingly.




Best Regards,


You must be Logged on to comment or reply to a post.
  • Hi Sujit – Do we have to request the operations team to increase the worked node size/Worker nodes in any type of subscription and will this incurr any additional cost ?

    Suppose if customer  opted for standard version(Also is it possible to check what version of HCI subscription customer has opted based on the HCI Tenant details) and if there is a drastic increase in the message volume we get any notification from the operations time to get increase the worker node/s/size or will the operations team automatically increase the size of worker node or the count of worker nodes.



    • Removed my earlier reply since I was factually incorrect.

      ———— Corrected reply  ———————-

      Hi Rajesh,

      With the standard edition of HANA Cloud Integration, you do not have to send a request to cloud operations and neither do you have to pay additional licenses for the extra VMs. You procure a standard license based on the number of connections and bandwidth. So, it is our responsibility for ensuring it is up and running as per the SLAs.When the load drastically increases, we shall increase the VMs accordingly so that the message processing goes through fine.

      Thanks and Regards,



  • I assume these worker nodes are spun up across availability Zones in same region? If so that is great.

    What is the recommended region failover practice? If lets say the whole region is down and disaster recovery is needed.