Proactively Monitor SAP Systems in the Cloud

⇨ SAP customers need to understand how their SAP and non-SAP systems can impact each other.

⇨ It is important for customers to have visibility across end-to-end business processes.

⇨ Visibility allows users to detect issues earlier, resolve problems more quickly and prevent negative impacts.

There’s no question that SAP customers have experienced a maturity journey in recent years. Having graduated from the days of traditional data centers running legacy technologies, most large enterprises today are running at least one cloud technology. Many are embarking on SAP cloud migrations or upgrade projects. While this move to the cloud reaps many benefits such as cost-savings, flexibility, reliability, high availability and data access improvements, it also enables more diverse and flexible landscapes than ever.

With hybrid environments being the norm these days, SAP customers are often running SAP SuccessFactors, SAP Ariba and SAP Business Technology Platform (BTP) or SAP Cloud Platform Integration (CPI) along with other non-SAP cloud solutions, while perhaps keeping certain systems such as for finance on-premise. Hybrid situations themselves can be complex, but what creates even more IT complications is that many customers are opting to maintain multi-cloud landscapes running a certain part of their cloud in Google Cloud while other integrations are going into Microsoft Azure or Amazon Web Services (AWS), for example.

If a customer is running its financials in SAP S/4HANA on-premise or in the cloud but doing its procurement with SAP Ariba, managing the interfaces between those applications becomes much more complex in a multi-cloud environment. There are many interdependent components to keep track of regardless of whether the business is moving to a public cloud provider, a hybrid cloud, or even SAP’s own cloud offerings.

Monitoring Complexities when Moving SAP HANA to Cloud Environments

To ensure continuous operations and avoid business disruptions, SAP customers need to understand how they’re on-premise or cloud SAP and non-SAP systems can affect each other. That is why the ability to successfully monitor complex hybrid and multi-cloud environments are so important for SAP customers today.

Moving to the cloud requires that a company’s business and IT teams monitor at a different level than before. Customers who have moved their business-critical SAP systems to the cloud want improved resiliency in a high-availability (HA) environment. But a more complex HA setup requires more monitoring. Rather than just monitoring a single set of servers for a primary production instance, HA deployments require monitoring the secondary instance and the Pacemaker cluster at the operating system (OS) level to ensure any outage results in a seamless failure to the HA instance.

Yet customers with complex environments like this still want to be able to easily monitor all the systems in their landscape and track key performance indicators (KPIs) from SAP’s application layer, integration layer, database, and OS layers — as well as KPIs from shared integrations, load balancers, firewalls, message queue services, web services, archive systems, and many other non-SAP applications. They want to connect their Google Cloud, AWS or Azure monitoring, and they want a seamless failover with no human intervention. Crucially important is that they want to respond to issues in a proactive rather than a reactive fashion.

When SAP customers upgrade their IT landscape to include SAP HANA and are running their enterprise applications on it (whether that be SAP Suite on HANA SoH, S/4HANA, BW/4HANA or something else), administrators and database managers have to expand their view and skill sets to support these applications. “SAP HANA is a very performant database, but it requires very specific tuning in order to operate properly,” said Joe Darlak, Head of SAP Solution Management at Google Cloud.

Not only does the performance of the database become critical but so does its relation to the application layer and the underlying infrastructure. “When you’re monitoring any SAP system, you’re not only looking inside the application and database layers for inefficiencies, memory swapping, short dumps, work process errors or batch job errors, but you also need to understand and address any constraints that infrastructure may have on the operation of the entire system to make it work correctly,” Darlak said.

SAP certifies compute and storage infrastructure for running SAP HANA and sets networking latency targets, so it’s essential to architect a solution that uses the right components that are optimal for the use case. According to Darlak, “With SAP HANA, the infrastructure, networking requirements, performance, and optimization all become more important because the database performs so well that other components may become the bottleneck. Understanding CPU capacity and memory utilization at the server level become essential. And it’s not just about tuning what you have. It’s about making sure you’ve got the right fit in terms of infrastructure and a path forward to handle future growth.”

Network or infrastructure architects will need to know how to navigate the hyperscaler’s cloud console. That might require those managing and provisioning the network to learn some new skills or take some cloud architect training to become familiar with the new functionality and agents that are native to the cloud provider. Google Cloud, for example, has an agent to deploy on every virtual machine (VM) that’s running an SAP system that collects data about the Google Cloud host machine and its environment and provides the data to the SAP Host Agent. It has another agent that specifically collects metrics from the SAP HANA database and feeds them to Cloud Monitoring so they can be displayed in the Google Cloud console.

“There’s data flowing both ways. Either way, you can pump different data sets from the cloud logs into the SAP system or metrics from the SAP system into Cloud Monitoring,” said Darlak. “But, of course, nobody wants to go through and parse a bunch of cloud logs, which is where a monitoring tool or a dashboard that picks these things out and displays them to you becomes valuable — a tool to weed through all the details.”

Viewing SAP Operations through a Single Pane of Glass

Jointly created by SoftwareONE and Splunk, SAP gives customers a platform that can correlate SAP and non-SAP data from disparate systems, load balances, middleware, firewalls, and cloud metrics — the entire landscape — and provide visibility across end-to-end business processes. The solution is provided by Splunk as part of its Observability solution portfolio. How it works is SoftwareONE’s PowerConnect collector lives inside SAP systems as an SAP-Certified Add-On. It extracts high-fidelity SAP data via 247 standard SAP function modules, reports and tables, and then filters, timestamps and obfuscates the data before sending the data to Splunk in real-time. Uni-directional from SAP to Splunk is secured via SSL, and the beauty of Splunk is that SAP data can then be correlated with any other non-SAP data sources in Splunk and Splunk Service Intelligence for SAP Solutions.

“We exchange an SSL certificate from the source SAP system to Splunk, and the handshake creates a fully encrypted link between the two environments,” said Gene Kaalsen, SAP Solution Strategist at Splunk. “PowerConnect comes preinstalled with 956 SAP use-case scenarios, and the out-of-the-box dashboards have inbuilt Machine Learning in place. It becomes even more valuable over time as it continues to learn how your SAP landscape behaves. The installation takes about 45 minutes, so the time to value is extremely fast.”

Kaalsen says, “Splunk Service Intelligence for SAP provides predictive service health scores, machine learning-driven insights and AI-directed troubleshooting. It comes out of the box with more than 450 adaptive threshold and static SAP KPIs, and its easy-to-build visualizations can track every important metric in a customer’s cloud migration or upgrade journey. For example, how an OS memory issue can ultimately affect certain background work processes and response time, resulting in short dumps.”

For native monitoring, Splunk supports monitoring of OS platforms such as Windows, UNIX and Linux, as well as business services (aka well-known package apps), which includes Microsoft 365, ServiceNow, salesforce.com, and, of course, SAP ABAP, Java and Cloud products.

“Because PowerConnect is installed inside the SAP solution, no matter where your SAP infrastructure is — on-premise, running on a hyperscale platform like GCP or AWS — it can grab the data and send it to Splunk,” said Pradeep Bhatt, SoftwareONE’s Global Product Manager for PowerConnect. “We have the visibility of whatever SAP has the visibility of.”

The overall benefits that Splunk offers SAP customers extend beyond providing end-to-end visibility of the environment’s SAP and non-SAP applications in one pane of glass. It also does:

• Eliminates siloed monitoring by consolidating monitoring tools into a single platform, thereby reducing cost and increasing revenue
• Provides near-real-time responses as opposed to waiting for batch updates, often the next day
• Allows for historical forensic analysis of data and easier audits since replicated data (not purged) has extended retention times in Splunk, which also saves disc space
• Offers dynamic security, handles custom tables / systems and is configurable for data collection and alert settings
• Delivers high-quality compatibility with SAP Products through 10 SAP-Certified integrations, including for SAP S/4HANA Cloud and RISE with SAP

According to Kaalsen, the beauty of Splunk is that customers can bring all their security, OS and VM metrics — basically everything that exists in their landscape — into Splunk and correlate any component or any KPI with that data. “The main purpose of our AIOps technology is to provide self-healing SAP landscapes where users could be notified 45 minutes before something might degrade and avoid an incident, based on previous teachings of sequences of events that resulted in a service degradation,” he said. “We prevent an outage based on Machine Learning and suggest actions to the On-Call team based on similar previous issues. We can automate responses by running a script, or a robotic process automation runbook. Some customers have reduced outages by 80%, and in other instances, we’ve avoided incidents three or four hours before they would have happened.”

He added that even for customers with very large and complex multi-cloud environments — from 200 to 2,000 SAP SIDS — have achieved full visibility across all those instances with Splunk without having to manually set individual KPIs in a monitoring tool. “Splunk Machine Learning and Artificial Intelligence identifies what is typical for any environment based on seasonality or recurring patterns, all without having to involve a data science team,” he said.

“Through the built-in machine learning, Splunk identifies a suspicious log-in activity, such as someone logging in at an anomalous time, or from an IP address they’ve not logged in from before or running a transaction they rarely run,” said Kaalsen. “And that real-time alert could go to the SAP security team, the centralized SOC or execute an automated runbook to lock that user in SAP.”

Providing Machine Learning-Driven Insights before, during and after Cloud Migrations

Because Splunk is vendor and cloud-service agnostic, it doesn’t matter which cloud console captures events and metrics, and logs data. Splunk brings all that data together, correlates it and draws predictive and event analytics and service insights with more than 2,000 integrations supported. The machine learning techniques employed are real-time clustering for events analytics and regression modeling for predictive analytics among other machine learning algorithms.

“We can correlate this data and give users insights that a certain application is currently showcasing degraded experience because a particular database service from a cloud provider is experiencing higher latency. That’s what AI-directed or machine-learning-driven troubleshooting is,” said Tapan Shah, Principal Product Manager – Observability for IT at Splunk. “We direct users to what to look for and where to look. We give them those insights up front and help them identify where to troubleshoot from. We point out their worst offenders and help answer the ‘So what?’ question for a problem. We tell the user the next thing they should be looking for, rather than giving them a pool of data and asking them to mine the data to get insights.”

The result is that customers can detect issues earlier, resolve problems more quickly, improve the mean time to repair (MTTR) and prevent the business and customers from a negative impact. Armed with analytics called “predictive service health scores,” companies can identify their perceived revenue loss if a problem were to persist for several hours. A score is applied to the IT system based on its overall health along with a forecast of the situation 30 to 45 minutes later. “Our customers commonly use these predictive service health scores to predict degradations in their environment before they occur,” Shah said.

A popular use case for Splunk Service Intelligence for SAP is monitoring end-to-end system performance before, during and after an SAP cloud migration project. Say an SAP customer is running SAP ERP or SAP HANA on-premise and it is migrating to a cloud-based solution, Splunk provides visibility into the before and after picture of the workloads running on the deployment, but it also helps the customer watch or observe that deployment in the phase of migration.

Before Migration: During the blueprint phase, the customer can use Splunk to baseline all source environment application, integration, database and system KPIs across security, observability and business processes. Customers can establish benchmarks and ensure visibility into every dependent component of the SAP system in the entire landscape.

During Migration: Splunk allows users to monitor the overall migration experience in the destination environment. It continues to monitor the system health on a configurable basis such as hourly or daily to ensure that all those dependent actions, workflows, backend processes and interfaces work as effectively as they should. Users continually monitor the impact on SAP system health with the help of deep-dive swimlane views to correlate KPIs to any changes in the infrastructure landscape, including Active Directory, firewalls, load balancers and message queue services.

After Migration: Customers use Splunk to remove any ambiguity and get data-based answers to questions like these: Did end-user performance improve post-migration? Are the number of failed interfaces and user logins trending up or down? Was the CPU and memory consumption as low as expected afterward?

“Quite a few of our customers have migrated from on-premise ECC to SAP S/4HANA or maybe another SAP software-as-a-service offering and are using Splunk to monitor their migration and compare the system performance before and after,” said Bhatt. “Because Splunk keeps the data for any length of time you want — for three, six, or twelve months — you can use it to monitor the SAP system before and after the cloud migration.”

Besides identifying system performance trends such as consumption changes, Splunk provides migration insights on end-user performance, capacity utilization from development to testing environments, and the number of failed user logins. It brings in all the vulnerabilities across the migration as they happen through approximately 130 dashboards for security, infrastructure, background processes, BPM IDOC to business object relationships, and network and database monitoring, which are immediately available upon installation.

Easy and Accessible Monitoring

When an SAP customer moves to the cloud, there are additional pieces of the environment that will need to be monitored, so a more comprehensive solution is necessary. Splunk brings in all the SAP and non-SAP data across the customer’s vast hybrid and multi-cloud environments and makes it easy for administrators to maintain the infrastructure. It helps customers make confident, data-driven SAP cloud migration decisions to speed up their business, fuel growth, and meet increasing demand.

“The more data you have, the more analytical you can be about determining proactively what issues may be occurring or may be about to occur,” said Darlak. “When customers are using Splunk and migrate to cloud, it’s just that much easier for them. Otherwise, they have to build up their monitoring with custom queries and dashboards. Why build something from scratch if there’s a tool that already does that for you and makes it easier on you? Splunk is one of those tools that just makes it simple.”

For more information on Splunk, visit their SAPinsider showcase page at: https://sapinsider.org/vendor-showcase/splunk/

For more information on SoftwareONE, visit their SAPinsider showcase page at: https://sapinsider.org/vendor-showcase/softwareone/