16.2.1 User requirements

Suppose that JP1/IM is to be deployed in the business system described in 16.2 Model for performance evaluation, based on the following user requirements (italics indicate JP1/IM keywords).

• A new monitoring server is installed to centrally monitor the business system, and investigate machines where failures occur as necessary.

  -> System hierarchy (IM configuration), Centralized monitoring using the Central Console, and event search

• The style of monitoring involves three operators, each assigned a specific role as in the existing system.

  -> Applying event receiver filters

• Monitoring targets are grouped according to the purpose of the business system.

  -> Using the Central Scope (implementing monitoring range settings)

• The operator is to be notified automatically when an error requiring immediate attention occurs, even if after hours.

  -> Using automated actions

• A history of day-to-day monitoring is kept.

  -> Saving event list information (CSV snapshot)

To satisfy these user requirements, the appropriate settings must be made in JP1/IM. For the purposes of this model, the requirements are satisfied by completing the following settings:

System hierarchy (IM configuration)

A new monitoring server is installed to centrally monitor the business system.

-> A system configuration consisting of one manager and 60 agents

Centralized monitoring using the Central Console

Events forwarded from agents to the monitoring server are suppressed under the following conditions:

• Events forwarded from a batch-processing server: High priority; forwarded when Warning or higher

  -> Assume 20 to 30 such events per server per day

• Events forwarded from a Web server: High priority; forwarded when Warning or higher

  -> Assume 10 to 20 such events per server per day

• Events forwarded from a PC: Low priority; forwarded when Error or higher

  -> Assume 1 or 2 such events per PC per day

Taking the higher of the two values, you can expect approximately 370 events to be forwarded each day, or 400 when you include the events that occur on the monitoring server (approximately 20 to 30 events with the event level Warning or higher). Add a further allowance of 100 to the total, and set the resulting value as the number of events to be stored by JP1/IM - Manager at any one time (the event buffer size).

However, supposing that you want to check past events at the same time as recent ones, set the event buffer to 1,000 so that JP1/IM - Manager stores two days' events.

Also, on the assumption that event searches will be carried out as needed from the three monitor machines, set the number of events acquired from machines where errors occurred (the number of events to acquire in one search) to 100.

Utilizing event receiver filters

Filter the monitored events according to your requirements. Then, calculate the number of events to be saved (the scroll buffer size) in the JP1/IM - View for each operator as follows:

• Operator A: Monitors batch-processing servers. Set the scroll buffer size to 500 events ((30 events × 7 servers + 40 event allowance) × 2 days).

• Operator B: Monitors Web servers. Set the scroll buffer size to 200 events ((20 events × 3 servers + 40 event allowance) × 2 days).

• System administrator: Monitors the entire system. Set the scroll buffer size to 1,000, the same as the event buffer size.

Using the Central Scope (implementing monitoring range settings)

Create the following monitoring tree:

The total number of monitoring nodes in the tree (the total number of monitoring nodes managed in JP1/IM - Manager) is 64, made up of four monitoring groups and 60 monitoring objects.

Implement monitoring range settings to restrict the number of monitoring nodes appearing in JP1/IM - View as follows:

• Operator A: One monitoring group, seven monitoring objects

• Operator B: One monitoring group, three monitoring objects

• System administrator: Four monitoring groups, 60 monitoring objects

Automated actions

To enable smooth error detection and basic troubleshooting, 20 automated actions are defined to suit the operating requirements. Of these, five are triggered by multiple conditions joined by an AND condition^#.

#: An event in an AND condition is kept in memory for the specified keep limit (60 minutes by default) or until the AND condition is satisfied. Depending on the situation, assume that in a 60-minute period approximately 30 events will be kept (an estimate obtained by dividing the number of events in a day (500) by 24 hours, and adding 10).

Saving event list information (CSV snapshot)

When monitoring is interrupted for some reason, such as at the end of the working day, a CSV snapshot is taken of the event information displayed in JP1/IM - View.

Other

Do not modify settings, such as the log file size, that increase the amount of disk space used. Continue to use the default settings.

The following figure shows the system operation after JP1/IM is deployed.

Figure 16‒2: Example of business system operation (after JP1/IM deployment)

Note

The memory required by JP1/IM to monitor the system takes two forms:

1. Memory required by JP1/IM on a constant basis

2. Memory allocated and released for specific purposes during JP1/IM operation

In the above figure, the relevant keywords for the first type of memory are the event buffer size and number of monitoring nodes for JP1/IM - Manager, and the scroll buffer size and number of (displayed) monitoring nodes for JP1/IM - View.

The relevant keywords for the second type are Number of events to acquire in one search and Number of AND events for JP1/IM - Manager, and Saving event list information (CSV snapshot) for JP1/IM - View.

When estimating the amount of memory required by JP1/IM, use the sum of the maximum amount of both types of memory.

Use of the Intelligent Integrated Management Base

If the system name is Business System and the subsystem names are Batch Process Monitoring, Web Server Monitoring, and QA Machine Monitoring, the operating status tree is created as shown in the following figure:

In the IM management nodes, the number of management groups will be 5 and the number of management objects will be 60.

Estimate your system requirements in terms of this framework. Based on this information, estimate the memory usage and disk space requirements of JP1/IM from the equations in the Release Notes for JP1/IM - Manager and JP1/IM - View.

From the results of the equations, make sure that JP1/IM deployment will leave some leeway in the resources of each machine. If you discover that a machine has insufficient memory or disk space, you must consider adding more memory or upgrading the disk. If this cannot be done, you might need to revise your requirements for the system in a way that accommodates your existing hardware.

Also, make sure that a sudden increase in the number of events acquired and processed by JP1/IM due to an unexpected error will not significantly impair performance in terms of displaying events and executing automated actions.

To Page Top