16.01.2013 Views

Microsoft Sharepoint Products and Technologies Resource Kit eBook

Microsoft Sharepoint Products and Technologies Resource Kit eBook

Microsoft Sharepoint Products and Technologies Resource Kit eBook

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

250 Part III: Planning <strong>and</strong> Deployment<br />

Healthy operation requires that you monitor each component <strong>and</strong> alert system<br />

administrators of problems. Most large enterprises already have policies in place for<br />

managing hardware, networks, <strong>and</strong> operating systems. The SharePoint Portal Server<br />

architecture does not introduce any new considerations with regard to those items.<br />

Monitoring <strong>and</strong> Alerting Requirements<br />

The more distributed an environment is, the more it requires monitoring. Failure to<br />

monitor the status of the architecture results in a reactive administrative model in<br />

which action is taken only after the problem has become serious enough for users<br />

to report it. The monitoring solution for this architecture provides a proactive administrative<br />

model in which problems are automatically detected before they escalate to<br />

a level that affects the system. To meet this aim, you should be aware of general<br />

monitoring principals <strong>and</strong> critical elements that need to be monitored.<br />

It is imperative that alerts be used to bring any failure to the immediate attention<br />

of the system administrator so that it can be rectified. If this is not done, the<br />

infrastructure can slowly decay until it affects the performance of the website.<br />

The monitoring <strong>and</strong> alerting infrastructure can also benefit scalability. Defining<br />

alerts based on system usage makes it possible to be proactive <strong>and</strong> start scaling the<br />

environment before users are affected. For example, an alert can be triggered when<br />

processor utilization on the Web servers is consistently above a preset limit. This can<br />

provide an indication that more servers or upgraded server hardware are required.<br />

The following topics discuss several fundamental monitoring principles.<br />

Monitoring Unobtrusively<br />

A poorly designed or poorly implemented monitoring strategy can adversely affect<br />

system performance or operation. In general, the monitoring solution should have<br />

as minimal an impact as possible on the system it is monitoring, while providing all<br />

relevant information.<br />

For example, when using a probe, it is sufficient to retrieve a single entry,<br />

which indicates that the server is functional, instead of retrieving multiple entries.<br />

The system should be probed only as often as absolutely necessary to provide adequate<br />

responsiveness. This limits the additional burden on the servers <strong>and</strong> applications.<br />

Probing a system too soon <strong>and</strong> too often returns failure data sooner, but it can<br />

place an obtrusive burden on the services.<br />

Avoiding Cascading Failures<br />

If a failure occurs, it can trigger other alerts in the monitoring solution. For example,<br />

if one set of servers running Internet Information Services (IIS) fails, this might place<br />

an additional load on the remaining servers, in turn generating alerts from those<br />

servers. If this happens, you should disable noncritical services or applications to<br />

reduce the load on the remaining servers. You should also evaluate system capabilities<br />

to proactively provide extra system capacity in case the problem recurs.

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!