The SAP EarlyWatch Alert service covers many topics regarding system health, performance and security. This blog describes how you can self-organize your consumption of the service to get the most benefit out of it: Stay informed about urgent topics by setting up notifications, schedule regular meetings to follow-up the comprehensive alert list and pro-actively seek for optimization regarding capacity planning and application optimization!
Here is your best practice schedule:
- On Demand: Follow-up the notification emails for urgent topics
- Once per month: Work your way through the alert list
- Once per quarter: Plan your hardware capacity
- Once per quarter: Tune your applications
- Before software update: Collect patches and notes
- After upgrade or conversion: Compare performance and find regressions
Before we come to the detailed descriptions: Set-up a meeting cadence for these tasks listed above in your calendar!
On Demand: Follow-up the notifications emails for urgent topics
You need an initial set-up for this task. Here are the steps:
- Open SAP EarlyWatch Alert Workspace
- Filter for the systems, you are responsible for
- Save your settings as a new “View” and give it a speaking name
- Go to the notification area and press “manage”
- Switch on the email notifications for SAP EarlyWatch Alert Workspace
- Click on the icon for “Customize in-app settings”
- Find your new view name and use the drop-down menu and select all scenarios
This way you will be informed on the following occasions:
- New decisive red alert (most important category of alerts)
- No current service report (Data from your system is missing since 2 weeks)
- SAP HANA table will reach the 2 Billion limit
- SAP HANA memory forecast is critical
- Severe slowdown happened
When you receive a notification details are provided in the email.
New decisive red alert:
Use the link in the email to navigate to the report chapter. Is the alert relevant for you? Who should be informed?
No current service report:
There are two links provided in the email: one to the latest available report, which is 2 weeks old. And one to the app SAP EarlyWatch Alert Status where you find the transmission and processing details in the backend at SAP. In case there were no download received at SAP, you need to troubleshoot on your side for on-prem systems. In case of a private cloud system operated by SAP ECS, open a service request. In case of a public cloud system at SAP, open a ticket to the SAP EarlyWatch Alert team on component SV-SCS-EWA.
SAP HANA table will reach the 2 Billion limit:
If you have enabled the notifications you will be notified as early as possible, when we detect a critical growth rate of a SAP HANA table. The predicted worst case date when the table could potentially hit the 2 Billion limit is calculated based on a time series of weekly data points. Depending on the selection of the time frame the result may vary. We have an automatic change point detection that can reset the history to improve the forecast result. Take a look at this blog to learn how to overrule the automatic result and adjust the forecast.
SAP HANA memory forecast is critical:
The SAP HANA memory forecast consists of two components: data growth and spikes from temporary memory usage, e.g. for delta merges or expensive SQL statements. Also in this forecast you can influence the data point selection and hence the worst case date manually. See also this blog. In case you find a critical growth trend, schedule a capacity planning meeting, see below for details.
Severe slowdown happened:
If you receive a severe slowdown notification, this indicates that end-users were effected by very long response times globally in your system in the last week. Use the link to navigate to the chart where you see the average response time per hour from Monday until Friday. Click on the left side chart to open another week for reference. Ask yourself: What happened during this hour? Which time component had increased the most, e.g. database time? Do you find a correlation to high memory or CPU consumption? Try to find the root cause and, if needed, get advice from SAP or an SAP Partner operating your system.
Once per month: Work your way through the alert list
There is no need to read the full SAP EarlyWatch Alert report page by page! Use the SAP Solution Finder instead. For more instructions read this blog. This app provides an aggregated list of alerts grouped by topic not by system. It allows you to drill down to the chapter level per system to read the details if needed.
Your task will be to filter for your team’s working list. You can filter e.g. for system IDs and alert categories, as well for only new alerts. Follow up the alerts and try to understand the risks and recommendations.
There will be a new feature introduced next year: You will be able to hide and snooze alerts on company level. This allows to put alerts on mute, in case they are not relevant right now. You will be able to add a comment for the reason to give feedback to SAP on alert level.
Once per quarter: Plan your hardware capacity
The SAP EarlyWatch Alert Dashboard provides a system specific overview on hardware consumption. SAP HANA memory and CPU consumption charts are accessible and provide a long-term history up to 4 years, depending on when you connected the system to SAP EarlyWatch Alert.
Use the SAP HANA memory area chart to drill down into the view where the different parts are color-coded: In blue you will find the column store representing your data growth. You can switch-on the data footprint forecast for the next 12 month, if you have enough history data.
Check the top 40 tables by memory size by clicking into the bar chart of the memory area chart. Compare the list with the situation 3 month before: Which are the top 5 tables in memory size? Do you see saving potential in data archiving, deletion or using SAP HANA native storage extension? You can also use the 2 Billion App on SAP EarlyWatch Alert Workspace to monitor table growth: Simple switch to “Time Frame” = “All objects – any date”. For further details use the Data Volume Management Dashboard in SAP ONE Support Launchpad.
More rarely the SAP HANA CPU consumption is an issue. Check out the “Analyze CPU utilization” app from the history chart of the CPU detail page. Here the applications are color-coded and you can drill-down to the SQL statements with their percentage of CPU usage.
Decide wisely on hardware up-sizing. Maybe you still can optimize your consumption. Add your knowledge of further project milestones bringing load and data increase to your considerations. Discuss your findings in your team on a regular basis, e.g. in a quarterly meeting.
Once per quarter: Tune your applications
Low-hanging fruits first:
- Use the SAP Solution Finder to search for missing SAP HANA indexes by filtering for alert category “Performance -> SQL Statements”.
- This category will also inform you on expensive SQL statements with a high overall system impact: Search for the alert title “Problematic SQL Statements” with a red alert rating. Forward the statement details to your development team.
Take your time to evaluate the performance of your business applications. From the SAP EarlyWatch Alert Dashboard you can drill-down into the “Top Transaction” details for GUI apps and into the “Top odata Requests” for Fiori apps. Select the top 10 entries to get the time chart on the right panel.
You want to find the following patterns:
- Upward trend of the response time for many apps: Correlation can indicate a system-wide bottleneck
- Upward trend of a single app: Is the data used in the app increasing significantly over time causing an expensive SQL statement to run longer and longer?
- After an upgrade: Response time shows step-wise increase: Look for SAP notes for a known issue or in case of custom code contact your developers
- Monthly pattern of increase and decrease: Does this make sense from the application side? Or is it a side effect from the overall system load?
Before software update: Collect patches and notes
A smart move to find recommended SAP notes and patches is to use the SAP Solution Finder. Here is a list of search terms which you should enter:
- “hotnews” – finds important notes with high risks relevant for your landscape. The prerequisites are already checked and for some issues the service identifies if data loss or inconsistencies have already happened.
- “side effect” – finds SAP notes for EWM, TM, PP/PM and FI
- “rtcctool” – finds all updates needed to ensure the latest recommendations
- “upgrade” – finds various tailored recommendations for upgrading software components
After upgrade or conversion: Compare performance and find regressions
The app Event Statistics in SAP EarlyWatch Alert Workspace automatically detects product upgrades, SAP HANA migrations and SAP S/4HANA conversions. The response times are compared automatically and visualized in a bar chart that represents the last 8 weeks after the software event. Find regressions in different task types and use the link to navigate to the SAP EarlyWatch Alert Dashboard to identify specific transactions or odata services which may be affected.
Make sure that you connect also your pre-production system to SAP EarlyWatch Alert Workspace to detect regressions earlier before they affect the prod system. The system needs to be sending data at least 8 weeks before the upgrade or conversion.For more details see this blog.
Stay curious!
Use the SAP EarlyWatch Alert Workspace to gain an overview on all apps. There is a What’s New? card which features new blogs regarding the service. I hope you manage to introduce some new habits into your tight schedule to stay informed and be more pro-active!
Feel free to share your own SAP EarlyWatch Alert Workspace best practices in the comments!
