SAP BusinessObjects Monitoring – Part 2

In Part 1, we looked at setting up the new Monitoring Application in SAP BusinessObjects 4.0. So, now that we have the application set up, in Part 2, we will look at using the application. Monitoring is a Flash application (What were they thinking?) built into the CMC. So, to get there, log into the CMC, and click Monitoring from the main menu. There were some substantial changes made to the Monitoring application in Support Pack 4 (SP04). This article will be based on SP04, but I will point out what is different from the BI4.0 base install.

Dashboard

There are five tabs across the top of the application. The first tab is the Dashboard, so that’s what you see when you first launch the application. The Dashboard has five panes (four prior to SP04):

  • Overall Health: A single icon gives you a quick overview of the health of the system. When you mouse over it, it looks like you can click to drill into it, but you can’t.
  • Recent Alerts: We will talk about Alerts later, but this pane shows any Alerts that were triggered within the past 24 hours, so you can get a quick view of the recent health of the system.
  • BI Landscape: We’ll talk more about Watches later, but this pane includes a graphical or tabular view of the Watches within the system. You can drill into any of the Watches shown here to get more information.
  • Watch Graph: The Watch Graph is directly below the Dashboard Watches pane, and displays a graph for the select Watch above. There are two views of each Watch. There is a Live view, which constantly updates with the status of the Watch, and there is a Historical view that allows you to view the status of that watch over a historical time period.
  • KPIs: This pane displays three Key Performance Indicators for the system. You can click on any of them to view more details for Running Jobs, Pending Jobs, or User Sessions.

Monitoring Dashboard BI4 SP04

Metrics

Metrics are individual pieces of information that contribute to the overall health indication of the system. They are items that are measured within the system. The system comes with over 250 metrics. In addition, you can create your own metrics, based on the built in metrics. A few examples of standard metrics would include the following:

  • Number of Defined Web Intelligence Reports (I believe they meant documents, not reports)
  • Current Number of Tasks (Web Intelligence Processing Server)
  • Completed Jobs (CMS)

What you do with these metrics is completely up to you. You can use them in Watches and Alerts, setting thresholds for Warning and Danger alerts. The list of Metrics is quite impressive, and can generally tell you almost anything you want to know about the system.

You can view the current, or historical, value of a metric by selecting the metric in the left pane, then adding it to the right pane.

Note that the default view is a live view, that is updated every 15 seconds. Click the Go to History button to view historical data. When viewing historical data, you can define the time frame by clicking the calendar buttons.

Watchlist

The Watchlist tab allows you to monitor the Watches in the system. What is a watch? No, it’s not that thing on your arm that tells you the current time. 🙂 A Watch allows you to set thresholds for the Metrics, and determine what values for a Metric represent healthy, warning, or danger levels. For example, You may decide that when the disk space for the Input FRS reaches 70% capacity, you want to trigger a warning, and when it reaches 85% capacity, you want it to trigger a danger status.

There are a number of built in Watches, which you will probably want to edit, changing the thresholds to levels that make sense for your environment. You can also decide which watches will be a KPI (Shown on the dashboard) and/or written to the trending database, allowing you to see historical data for that watch.

If a Watch evaluates to True, what action do you want to see happen? This is called the Notification. And how long should the Watch evaluate to True before that action is taken? This is called the Throttle. For example, if the number of concurrent running jobs exceeds X, should the system send a notification immediately, or wait until it sustains that level for X number of minutes? These are options that you can set within a Watch.

Some of the options on watches include sending an email, or triggering a Probe. You can have an email sent in case of a Caution alert, or Danger alert, or both. To do this, click the Directory button next to the Notification Settings.

Probes

A probe is a device that executes a workflow, mimicking the actions of a user. The purpose of a probe is to test user actions, and make sure that they work as expected. There are nine built in Probes. Some of them can be run as is. Some of them will need additional setup before they can be run. For example, the CMS Ping Probe, which simply checks to see if the CMS is up and running, can be run as is. However, the Interactive Analysis Probe (Which you should rename as Web Intelligence Probe) needs to have the CUID of a Web Intelligence document, so it can use that document to test if a document can be opened, refreshed, and saved in various formats.

You will need to modify the URL of the BI Launchpad Probe. Here is the correct URL:

http://<servername&gt;:8080/BOE/portal (Make sure you replace <servername> with the name of your server)

For the Web Intelligence Probe, I would recommend using one of the Sample documents, built against the eFashion universe. This way, the probe doesn’t hit your production database, and the document runs quickly. Remember, the purpose of this probe is simply to make sure that users can open and refresh Web Intelligence documents.

The system allows you to register new Probes, either Java based, or Script based. You can also delete existing Probes that you don’t need.

Alerts

All those Alerts that you set up in the Watchlist, will show up here. This is simply a list of all alerts that have been raised within the system. You can click on the alert to see the details of the alert. Keep in mind that alerts can also be sent to email addresses. For example, you can set up an alert that emails you if it crosses one or both thresholds. This is set up in the Watchlist tab.

Conclusion

Using the Monitoring application can be a challenging task, given the number of built in Metrics, and lack of training available from SAP. But don’t give up. It’s a powerful tool, with a lot of potential. If there is something that you want to do, and can’t figure it out, comment here, and I’ll see if I can help.
In part 3, we’ll discuss building a universe on the Monitoring trend data, so you can do historical reporting on the system’s health.

Advertisements

35 Responses to SAP BusinessObjects Monitoring – Part 2

  1. Ana says:

    Great information. Can’t wait to see more about the “Monitoring” universe.

  2. Chandu says:

    Thanks for insight information about new BO admin tasks.

  3. Gunther says:

    Hi, anyone having slow performance when trying to open one of the bigger watches like core health ? For me it takes a long time to load, anyhting I can do to speed this up ?

    • Swati Bajaj says:

      Hi

      Even I am facing a similar problem. I was able to create the desired watch but now when i want to edit that watch, it hangs and does not open for a very long period.
      Can someone suggest about what can be done to optimize such a situation?

      Thanks
      Swati

  4. Jens says:

    I am not able to set up a user to an already exisiting watch. A NullPointerException is thrown at any time :(.

  5. Chris Beck says:

    Hi Michael, great information as always.

    I would like to create an email alert that is received when a schedule is successful or failure. *Not the schedule itself, just a simple email saying “The schedule failed” or something like that. I have tried creating a scheduling system event for this, but cannot get it to work as I would expect. Basically, I never get the email as I expect. There is not an issue with the smtp configuration or anything like that, I presume it simply user error. Perhaps you can help!

    Thanks,

    Chris B

  6. john lobo says:

    Michael,
    Great information, This helps people like me to learn more about business object, BTW is there any metrics to find the maximum number of concurrent user in a day?

    Thanks

    John

  7. Pingback: SAP BusinessObjects Monitoring – Part 3 « Michael’s BI Safari

  8. bernhard says:

    hi,
    thx for your efforts to document monitoring BO, which i had expected to get from SAP BO 😉
    i am tired to google one metric i instantly find in default watches. it is “Health State” and obviously has the Value of 2 when the health state is fine. but what in detail is the state of 0 or 1 or is there another state?
    In general, is there a metric documentation available anywhere that helps me defining appropriate watches (f.i. is there a limit of current connections until the job crashes in publication processing, or so …)
    thank you very much!

    bernhard

    • Hi, bernhard. A metric can have 3 states. These are healthy, warning, or danger. These are defined in the Watches, and you can set the levels at which each state is reached.

      • bernhard says:

        Hi Michael,

        thank you, i was thinking that these 3 values mean essentially “error” = 0, “warning”=1 and “i’m fine”=0.

        I am someone who wants to understand the system background, so a “Health” can mean anything possible for me. I am curious: under which circumstances does the health state change? Is that a composition of all watches defined for a server that means if one of the watches raises a warning alert, the health state is “warning” as well? Or something else?

        I am a little frustrated because i didn’t find ANY valuable documentation of BO monitoring so far that assists me in defining good watches or understanding the system and their metrics.

        so if someone has a piece of doc found, could you share that with me? Thank you! 🙂

      • Yes, if one watch enters the warning state, then the overall system health will be warning.

  9. Thomas says:

    Hi Michael,

    When I switch to ‘Go to History’, it always show ‘NO DATA AVAILABLE’, do you know what I should do to see the historical data?

    Many thanks,
    Thomas

    • Hi, Thomas;

      Not all watches are written to the trending database. That’s one of the options on a Watch. If you edit the watch, make sure that is being written to the trending database if you need to see historical data for it.

  10. MDAS says:

    Hi Michael,
    I have a quick question as digging around forums etc, I’ve yet to get a clear answer.

    What formula would you use in a metric.. to show how much memory an MDAS server is currently using?

    Any advice much appreciated.

    • Off the top of my head, I’m not sure what that formula would look like. You might be able to find an existing metric that shows that number. I don’t durrently have access to a 4.0 CMC to dig into it.

  11. jonL says:

    Hello Michael, being new to the Montoring feature I found your blogs very helpful.

    Would you or anyone reading my post have a list of baseline metrics and thresholds that they have set up to monitor their environments for performance tuning.

    We have experienced what we feel is slow performnace within the BI Launch pad environment but we have no baseline to compare against. We are conductiong our tests on a new 4.0 SP7 system with only one user logged on.

    Any advice will be helpful
    Thank you

  12. SM says:

    Hello Michael

    Can you use Interactive Analysis Probe (Which you renamed as Web Intelligence Probe) to use CUID of a BO Dashboard. I tried, but it failed. If not, how can you create a probe to test a BO dashboard?

    Thank you.
    SM.

    • Hello, SM. I haven’t tried to create a probe to test a dashboard, so I don’t know if it’s possible. I suggest posting that question on BOB, to see if anyone else has tried and succeeded.

  13. jooken says:

    Hey,
    i created a new watch in the Monitoring Application. Is it possible to see this custom Watch on the Graphical View of the BI Landscape at the Dashboard?

  14. jooken says:

    This option i found already. But with this option i just can sie the custom Watch in the KPI Status box. Is it not possible to add a custom Watch to the Graphical View of BI Landscape??

  15. jooken says:

    I just have the default settings. But anyway, thank you very much for your help

  16. Rik says:

    great documentation. Thanks Michael !

  17. Hello Michael,

    Thanks for your post on Monitoring. Very useful!!!

    We are currently on BI4.1 and would like to use the Monitoring tool in order to keep a check on the audit log files.
    We recently ran in to an issue where the audit log files were not written in to the audit database. Due to this we lost around a weeks of audit data. We could not get them written to the audit db.

    To avoid such trouble again in the future, we would like to keep a watch on the audit log directory.
    I read in your blog post that we can set a watch on the consumption of the FRS and raise notifications if they reach thresholds.
    On the similar lines, is it possible to use Monitoring watch(or other) to see if there any audit back logs without being written to the database.

    Any help on this would really help us a lot.

    Thanks.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: