Deployment Automation and Monitoring in SLES for SAP Applications

SUSE will soon include a significant set of new initiatives and features into the SLES for SAP Applications product.

For a couple of months, SUSE has been working together with Microsoft Azure teams in development and advancing an open source project around SAP deployment automation and SAP monitoring.

Starting from the idea to modernize and simplify SAP HANA deployments, SUSE engineering worked on rewriting the deployment wizards in SALT, the SUSE automation standard that is also supported by SUSE Manager for any sort of automation. The goal of the project was to clearly standardize deployments and allow customers to use one level of tooling in various ways – from a Command Line interface, through some GUI driven process and SUSE Manager (the deployment could even be integrated and leveraged from other automation frameworks such as Puppet and Ansible).

With the goal to fully automate the major deployment tasks for SAP HANA, the team is a now at a point to be able to deliver a fully automated installation of a single node HANA system as well as a fully configured HANA Cluster deployment based on SUSE’s best practice documentation for Scale-Up SAP HANA cluster. With a integration with Terraform, even the environment can be automatically created on a KVM server, as well as the major cloud provider platforms to allow a simple and quick creation of additional database instances.

The team has also been focusing on NetWeaver deployment in the same manner, then next step will be gluing components together for a full SAP S/4 HANA deployment.

The project has been open source since the beginning and we obviously invite our users and partners to take a look and contribute at https://github.com/SUSE/saphanabootstrap-formula and  https://github.com/SUSE/sapnwbootstrap-formula .

While we were working on the deployment, another opportunity arose in our collaboration with Partners – OpenSource based monitoring of SAP HANA and the Cluster. 

In the container space, two tools are dominating the space – Prometheus and Grafana. Why not use them in the SAP space as well? 

To make a long story short, we looked into the software and came up with a few quick wins that we included into the deployment project for automated setup as well. 

At this point, the deployment automation can be configured to also setup a monitoring environment where each SAP HANA node automatically gets a Prometheus client as well as a HANA Exporter deployed. This provides key system metrics to any monitoring solution that can parse and understand the Prometheus format described at https://prometheus.io/docs/concepts/data_model/. This allows major monitoring solutions in the cloud or on premise to easily consume the data and include that into the monitoring capabilities without the need to install any additional agents on the Linux Servers.

At the same time, the metrics can also be consumed by a Prometheus server and rendered in a Grafana dashboard, for example, within SUSE Manager. The intention is to provide an integrated monitoring solution this way with the combination of SUSE Manager and SUSE Linux Enterprise Server for SAP Applications, but following the spirit of making a difference, clients can choose how much of the components they want to embrace, or if it is a better choice to integrate components into an existing solution.

In general, also for monitoring SUSE is looking into expanding the concept to NetWeaver and a full SAP S/4 HANA stack in the next few months. The code is part of the same github repository and fully open source. We do see huge interest right now from customers and partners, so check the solution out and join the project.

To encourage everyone, this project has already been used by a customer in production recently in Microsoft Azure.