AWS for SAP

End-to-End Observability for SAP on AWS : Part 1 Overview

Background

Customers are always looking to improve their operational excellence and resiliency of their core business processes in SAP systems. To achieve this, they need a consolidated monitoring/observability dashboard for their SAP landscape, so they can correlate issues, and understand whether they are in the database, application server, presentation, or network layers, (including Internet connectivity) or others.

AWS provides end-to-end observability through Amazon CloudWatch Application Insights for SAP HANA, CloudWatch Application Insights for SAP NetWeaver, CloudWatch Real User Monitoring, AWS Network Manager, Compute Optimizer, and CloudWatch Internet Monitor.

When these observability capabilities combined, we will be able :

  • To provide holistic root cause analysis of SAP systems, thus reducing MTTR (Mean-Time-To-Recovery) from days to just hours (if not minutes).
  • To enable proactive alerting to manage SAP systems before it fails for the SAP users.
  • To provide capacity forecasting and planning for SAP systems so customer can understand the key resources that may need to be right-sized to support their critical business processes.

SAP Architecture Layers

SAP ERP or S/4HANA systems are typically deployed in three-tier architecture as described below. This helps us to understand how the system behaves so we can observe it efficiently and effectively to ensure that you have a well architected system on AWS.

Figure 1. SAP architecture layers and AWS services that supports observability.

Presentation layer contain systems capable of providing a graphical interface for end-user to perform their activities

  • Presentation layer is also known as client layer, where user interaction happens with SAP systems.
  • For SAP User interaction purposes, we use SAPGUI (fat client installed on desktop) and/or SAP Fiori (HTML5 client that runs on desktop, tablet and mobile devices).

Application layer includes servers that runs application logic of SAP Systems

  • SAP application programs (ABAP) are executed in application layer.
  • Application layer serves as the middleware between the presentation and database layers.
  • Application layer is where the SAP dispatcher distributes the work load to the different work processes.

Database layer contains servers that store data that is required to run application logic of SAP Systems

  • Data store can be master data, business data, configurations, and ABAP programs.
  • Examples − SAP HANA, Oracle, Microsoft SQL Server, IBM Db2, SAP ASE, etc.

To run efficiently and effectively critical business processes in SAP, these various layers must work in tandem with no bottlenecks or issues encountered. This is why observability at every SAP layer is very important to turn a reactive problem troubleshooting to a pro-active system management to prevent business downtime and outage that introduce significant loss to business users.

Observability in AWS for SAP

Based on the SAP architecture layers explained above, AWS has developed the following services that provide end-to-end observability of SAP systems on AWS to enable you to effectively monitor your SAP workload moving from being reactive to proactive management enabling higher resiliency to support your business critical processes.

  • CloudWatch Application Insights helps you monitor your applications that use Amazon EC2 instances along with other application resources supporting SAP Systems.
    • When you use SAP HANA database to store business application data. You can find out more through this tutorial, and this blog. When you use SAP ASE database, you can find out more through this tutorial.
    • For SAP NetWeaver applications, which runs the application logic. You can find out more through this tutorial and this blog.
  • CloudWatch Real User Monitoring (RUM). It is part of CloudWatch’s digital experience monitoring which provides near real-time data on client-side application behaviour. It helps application developers and DevOps engineers to quickly identify and debug a range of potential issues, thereby reducing mean time to resolve (MTTR) and improving user’s experience in using SAP systems. You can find out more through this blog.
  • CloudWatch Internet Monitor. It provides visibility into how Internet issues impact the performance and availability between your SAP applications hosted on AWS and your mobile workforce.
  • AWS Network Manager. It provides tools and features to help you manage and monitor your network on AWS, which supports your business critical SAP systems. Network Manager makes it easier to perform connectivity management, network monitoring and troubleshooting, IP management, and network security and governance.
    • AWS Cloud WAN, is a managed wide-area networking (WAN) service that you can use to build, manage, and monitor a unified global network that connects resources running across your cloud and on-premises environments.
    • Infrastructure Performance, which allows you to obtain near real-time and historical network latency across AWS Regions and across or within Availability Zones for a specified time period.
    • AWS Global Networks for Transit Gateways, which enables you to create one or more global networks and then centrally manage those global networks across AWS accounts, Regions, and on-premises locations.
  • AWS Compute Optimizer. It analyzes the configuration and utilisation metrics of your AWS resources. When you combine with CloudWatch Application Insights for SAP HANA and SAP NetWeaver, Compute Optimizer will be able to report whether the resources are optimal, and generates optimization recommendations to reduce the cost and improve the performance of your SAP workloads.

We will discuss each of the services and their applicability to SAP workload in much more details in the upcoming blog series, please stay tune.

Use Cases

Let us now take a look at an example use cases on how all the above works to help SAP customers monitor their SAP Systems on AWS.

Use case where mobile users reported slowness when accessing SAP Fiori Launchpad from Internet

Figure 2. Use case where mobile users reported slowness when accessing SAP Fiori Launchpad from Internet.

We recommend to perform root cause analysis (RCA) process as below:

  1. With CloudWatch RUM, you can take a look at what the user is doing, where the user’s location is, what kind of performance that they are experiencing, and if there is any error encountered in the mobile device.
  2. Using CloudWatch Internet Monitor, you can see if there is any potential ISP connectivity issues that are happening around the user’s vicinity around the same timing which may impact user’s experience.
  3. You can see CloudWatch metrics for Application Load Balancer to see if there is any unhealthy state or slow response time detected at the internet-facing Application Load Balancer
  4. At the SAP Web Dispatcher, you may want to look at the general health of the EC2 instances using CloudWatch for EC2 Metrics, to see if there is potential bottleneck or issues at CPU, RAM, Storage, and so on
  5. At the SAP Application Servers, you can use CloudWatch Application Insights for SAP NetWeaver to see the key metrics such as Availability, front end response times, detected problems, pacemaker HA metrics for ASCS/ERS and others.
  6. At the SAP HANA Database layer, you can use CloudWatch Application Insights for SAP HANA to see the key metrics such as HANA out of memory situation, disk full situation, disk write queue, and others.
  7. At the Network Level, you may want to use AWS Network Manager to see if there is network level issues between Availability Zones (such as Latency measurement between application server and database server across multi AZ)

Lastly, You can also perform capacity analysis using AWS Compute Optimizer to see if any of the components (i.e. SAP Web Dispatcher, Application Servers, HANA Database) will need to be right-sized:

  1. When it is under-sized, for example : slowness due to continuous high CPU, RAM, Disk, and I/O utilisation.
  2. When it is over-sized, for example : continuous low utilization of CPU, RAM, Disk and I/O utilization.

Please note that the recommendations from AWS Compute Optimizer will need to be adjusted to match SAP certified instance types.

The above RCA process is an example on how you can leverage various AWS services and features to achieve end-to-end observability for SAP on AWS. It is possible to customise them into a consolidated dashboard for you to analyse faster, and to generate alerts using CloudWatch Alarms or you can also integrate it with Amazon Eventbridge so you can turn this into proactive monitoring to even do predictive-maintenance and create self-healing mechanisms before the user reported any error situations.

Conclusion

Observability is very important to understand your SAP workload behavior, and how your users experience using SAP systems while executing their business critical processes. A highly performant and resilient system will help with increasing your users’ productivity, freeing up time for innovation and value-adding activities.

With various AWS Observability services such as CloudWatch App Insight for SAP HANA and SAP NetWeaver, CloudWatch Internet Monitor, AWS Network Manager and AWS Compute Optimizer, you are empowered to do pro-active management and modernize your SAP workload.

You can find out more about SAP on AWS, Amazon CloudWatch Application Insights, Amazon CloudWatch Real User Monitoring, AWS Network Manager and AWS Compute Optimizer from the AWS product documentation.

Credits

I would like to thank the following team members for their contributions: Ambarish Satarkar, Ashish Tak, Derek Ewell, Mohit Biyani, Wong Whye Loong, Vijay Sitaram, Sriram Sabesan, Ramkrishna Borhade, Somckit Khemmanivanh, and Spencer Martenson.