Storage of Capacity Management data

Capacity and performance trends

Handling capacity- or performance-related incidents

If an alert is triggered, or an incident is raised at the Service Desk, caused by a current or ongoing Capacity or Performance Management problem, Capacity Management must become involved to identify the cause and find a resolution. Working together with appropriate technical support groups, and alongside Problem Management, all necessary investigations must be performed to detect exactly what has gone wrong and what is needed to correct the situation.

It may be necessary to switch to more detailed monitoring during the investigation phase to determine the exact cause. Monitoring is often set at a ‘background’ level during normal circumstances due to the large amount of data that can be generated and to avoid placing too high a burden on the IT Infrastructure – but when specific difficulties are being investigated more detailed monitoring may be needed to pinpoint the exact cause.

When a solution, or potential solution, has been found, any changes necessary to resolve the problem must be approved via formal Change Management prior to implementation. If the fault is causing serious disruption and an urgent resolution is needed, the urgent change process should be used. It is very important that no ‘ tuning ’ takes place without submission through Change Management, as even apparently small adjustments can often have very large cumulative effects – sometimes across the entire IT Infrastructure.

Capacity Management has a role to play in identifying any capacity or performance trends as they become discernible. Further details of actions needed to address such trends are included in the Continual Service Improvement publication.

Large amounts of data are usually generated through capacity and performance monitoring. Monitoring of meters and tables of just a few Kbytes each can quickly grown into huge files if many component s are being monitored at relatively short intervals. Another problem with very short-term monitoring is that it is not possible to gather meaningful information without looking over a longer period. For example, a single snapshot of a CPU will show the device to be either ‘busy’ or ‘idle’ – but a summary over, say, a 5-minute period will show the average utilization level over that period, which is a much more meaningful measure of whether the device is able to work comfortably, or whether potential performance problem s are likely to occur.

In any organization it is likely that the monitoring tools used will vary greatly – with a combination of system -specific tools, many of them part of the basic operating system, and specialist monitoring tools being used. In order to coordinate the data being generated and allow the retention of meaningful data for analysis and trending purposes, some form of central repository for holding this summary data is needed: a Capacity Management Information System (CMIS).

The format, location and design of such a database should be planned and implemented in advance – see the Service Design publication for further details – but there will be some operational aspects to handle, such as database housekeeping and backup s.

Дата добавления: 2014-12-23