No matter what platforms are at work in your landscape, you inevitably
have some business-critical applications where all hosts must be monitored
to achieve high availability and performance. These are often some of
the most complex systems in the company, encompassing multiple components
and platforms. A gradual reduction in performance or a sudden breakdown
of a component could appreciably affect the productivity of the entire
scenario and your sales, production, or service teams. And the
administrator is responsible for checking these all-important functions
to prevent problems before they can affect your business.
Take SAP Internet Sales as one example.
This application has been designed to give the high performance and fast
response times you need for an effective e-commerce website. Through Internet
Sales, your customers can configure and buy products from your online
catalogs. To prevent bottlenecks and to spread the workload, the solution
typically runs on a distributed software platform. Each server, middleware,
or backend system is scalable and can consist of several hosts to increase
performance. Figure 1 gives you a glimpse of the complexity of
a (CRM) Internet Sales scenario.
||A CRM Internet Sales Scenario
There is simply not enough time for an
administrator to log on to each of these host components to check their
status, so how do you effectively monitor such a complex and business-critical
system? The CCMS Monitoring Infrastructure, part of the SAP Web Application
Server 6.10, offers a central overview, the central monitoring system
(or CEN), for monitoring and managing the entire mySAP.com e-business
platform from a single console, without logging on to each and every component.
In CEN you can set up monitors to get, first, a central overview of your
complex server landscape (such as system availability, transaction-specific
response times, and so on), and then, if an error occurs, you can jump
from your central view right into detailed analysis. Agents are used to
collect the monitoring data of the different components, adapt CCMS central
monitoring to the heterogeneity of modern system landscapes, and offer
For landscape monitoring, you should use
at least three different monitors.2 First, you would use a
monitor to check that your systems are available. A second monitor checks
the general status of the available systems; here basic values, such as
syslog data, should be reported for each system. Last but not least, there
should be a monitor for checking the specific data of the systems, such
as shopping baskets and purchase orders for a CRM system.
Sticking with Internet Sales as an example, lets have a closer
look at availability monitoring. As you can see back in Figure 1, there
are quite a few components involved in the Internet Sales process.
If you look more closely at the CRM Internet
Sales component, which customers use to configure and buy products (see
Figure 2), the customer first searches for a product in catalogs
that have been exported from the CRM system to an external index server
(IMS) to accelerate access to the product data.
||Using CEN to Check the CRM Internet Sales
With the Internet and Pricing Configurator
(IPC), prices are calculated automatically during product configuration
with respect to customer-specific conditions, surcharges, or discounts.
When the product is ordered, an availability check is executed in SAP
Advanced Planner and Optimizer (APO) to make sure the product can be delivered
in time. mySAP Business Intelligence (BI) analyzes the data of the operative
CRM processes. The Internet orders themselves are processed directly in
the SAP R/3 OLTP System.
The administrators basic task is
to make sure that all of these interacting components are available. The
best way is to monitor whether the remote SAP systems and their application
servers are available for work using the CENs Alert Monitor. Availability
monitoring uses the alert and display functions of the monitoring architecture.
However, to determine system availability, monitoring uses the CCMS agent
CCMSPING. This agent is a standalone program that runs outside the SAP
system. By using an agent, you can efficiently monitor the availability
of a large number of systems from the central location (CEN), shown in
For the Internet Sales availability check,
it is not enough to check just the SAP systems; you must also check
the availability of web components such as ITS, BC, and
J2EE and processes, such as the CCMS agent. So its
recommended that you create three different monitoring tree substructures,
as shown in Figure 3:
- SAP Web AS
- Single Processes
||Internet Sales Availability Monitor and
the Substructures for SAP Web AS, WEB,
and Single Processes
In addition to the availability check using
agents, you can now use the Generic Request and Message Generator (GRMG),
shown in the WEB substructure, to check availability not only at the server
level but for components as well. For example, while CCMSPING merely informs
you whether the SAP Web AS is available, with GRMG you can also find out
if the web pages can be called correctly.3
Checking General System Conditions
After you have checked the availability of
the system, you might want an overview of
the general data for the systems (see Figure
General data consists of values that are important
for each system, irrespective of its function,
such as backup and syslog data as well as background
processing and update errors.
||Monitor for Checking General System Data
You might ask, Why not combine this
general monitoring data with the availability data into a single monitor?
There is a very practical reason: speed. It is possible to create a monitor
containing all of the data, however, the less data in one monitor,
the faster you can open it. Just imagine if you had a single monitor containing
all the data from both Figure 3 and Figure 4. Just to have a quick
look at the availability values, you would have to wait until all the
other data is loaded, too. So as a general rule, its a good idea
to make sure that your monitors do not consist of too many subtrees.
Checking System- or Component-Specific Data
Specific data consists of values
that are only relevant for a particular system
or component. In the subtrees of Figure
specific data is shown for CRM (B2B), LiveCache (APO),
BI, and Communication. Lets
look at the different subtrees.
||Monitor Containing System- or Component-Specific
The red entry in the first subtree, CRM,
indicates that an alert has occurred concerning the shopping basket. The
problem underlying this alert should be solved as quickly as possible
so that users wont experience delays in ordering. As outlined in
a previous article,4 you do not need to check this monitor repeatedly
to find out if an alert has been triggered. If you have assigned an auto-reaction
method to this Monitoring Tree Element (MTE), the system checks the monitoring
architecture for you, looking for alerts at regular intervals. If it detects
alerts, the assigned auto-reaction method will be started, and will send
an alert message directly to your pager or email.
As of SAP Web Application Server 6.10,
you can define central auto-reaction methods,
which are not started in the system where the
alert occurs, but in CEN. Thus, from a central
point, you can react automatically to events
that occur in the monitored components. And
if you have also assigned an analysis method
to the MTE, you only have to select the alert,
choose the wrench icon in the toolbar, and
you are linked to the right solution to check
the corresponding system.
Other specific data is shown for APOs
liveCache. As you can see in Figure 5, a yellow alert (warning) has been
triggered for the memory.
As shown in the third subtree, the CCMS
Monitoring Infrastructure now also offers monitoring of BI process chains,
such as the upload of the purchase order. Process chains are a sequence
of background jobs, whose relationships and status can be queried through
the maintenance view. To see this view, simply select the MTE DESCRIPTION,
click on the wrench icon, and you jump directly to the Process Chain Maintenance
view, where you can check on background processes (Figure 6).
||BI Process Chain Maintenance
As for the Communication subtree,
in a complex system landscape, consistent, up-to-date data must be guaranteed
in the local databases of the various systems. This is achieved through
replication mechanisms. Replication between the systems works using the
remote function calls (RFCs) tRFC or qRFC.5 Both RFC calls
can be monitored by CCMS. Because a communication problem could lead to
data inconsistency, this needs to be investigated immediately by double-clicking
the alert and opening the corresponding analysis tool.
With the CCMS Monitoring Infrastructure you can gain a big-picture look
at components, whether theyre in a single system or span multiple
platforms or components.
With the CEN system, CCMS provides a centralized
overview of the broad spectrum of systems and components to be monitored.
You can choose what part of the landscape is to be checked by a particular
monitor. If you have assigned the corresponding methods to your MTEs,
the monitors do the work for you by checking the environment and informing
you by pager, email, or fax when an alert has occurred. With the corresponding
analysis method assignment, you just have to double-click the alert to
get to the right analysis tool. And because the CCMS Monitoring Infrastructure
is an integral part of the SAP Web Application Server, the tools for increased
responsiveness and efficiency of business processes are already there,
right in your system.
For More Information
To learn more about CCMS and monitoring, SAP customers can log on to
or write to CCMS@sap.com.