Monitor SAP CPI system performance

Many of the customers I talk with start to see the SAP Cloud Platform Integration(CPI) as a critical tool for their integration needs. In that perspective, it makes sense to be proactive and monitor how it performances.

There are two pages that you can use to check the status reported by SAP as I write about on my blog it is the pages Cloud Status and Status page. They may be relevant to monitor both to warn about what is happening in your landscape.

But what about how to monitor you own tenant and how it is performing. That is the question that most people want to answer because it is what matters for them and then sooner they know about an issue they can setup manual processing or figure out how to do something else. It also provides a way to to messuare if SAP is living up to the SLA of CPI.

How you can messure your CPI tennant

You can setup logging in a number of different tool that will give you some indication of how the system is responding. You can also use the Figaf IRT tool that now will allow you to monitor your CPI system. It is pretty simple and does not require any coding form your side.

You will have a dashboard that looks like the following to see the performance of your system. You can see on the diagram that the system is down for maintenance for a period of time over the weekend. It is possible to see we cannot access the management and runtime engine for a period of time. This is a partner tenant so there may be a different performance and number of nodes as a real system. The system also calls an integration flow every 5 minutes to measure how the runtime is perfroming. This is then returning a response, but from time to time it takes almost 2 second to perform. Also, notice that during the maintenance period we still get a response in the range 400 ms. If it is possible to call process messages even when there is a maintenance window it is pretty cool.

Current view of the SAP CPI monitoring

We are now collecting all the metrics for this. Next is that we will be setting up alerts on this with our alert engine. Then you can get a notification if you are getting latencies above, i.e. 2 seconds or the CPU load is above 80%. Or create a Jira ticket with the simple webhook. We will also be adding status of your JMS queue and other relevant metrics asked for by users.

We have monitoring packages starting at only 150 EUR/Month and then you can also monitor other iflows.

Try it out at IRTCloud or you can also try it on your own deployment of the IRT server you can run on your own cloud. It also has quite a lot of other options to monitor SAP CPI both trigger alerts for failed messages, restart CPI messages and give users access to monitor data them self.

0 replies

Leave a Reply

Want to join the discussion?
Feel free to contribute!

Leave a Reply

Your email address will not be published. Required fields are marked *