Management & Control Services (MCS) can record and log system events and trace data that are generated either by MCS itself, or by any MCS-managed product or component in which event or trace logging is supported. MCS provides both event logging and trace logging services, which record the same types of events. You use the services for different purposes:
Events can be used to store and monitor a running log of events that occur within your system. The event service can generally run unattended, while allowing you to view, filter, and search for events as needed. In addition, you can create event notifications, so that the event system sends a message when specified event criteria are met.
Each MCS-managed product can log its own events in the MCS event log; the types of events that a product can log are specified by that product.
In contrast to event logging, you create custom trace configurations as needed to help track and troubleshoot any problems that occur. Trace logging is not enabled by default. To enable tracing, you must configure a trace for a particular product, and then start the configured trace. For instructions, see Configuring a New Trace and Starting or Stopping a Trace, respectively.
The type of information that can be traced for a product is specified by that product. Each product logs its own trace entries locally, and reports that data to MCS only when requested. Using MCS, you can view and save trace data from individual products in one central location.
Traces can be limited to a particular component, or they can be configured to follow a task across a series of clusters or servers. When you are finished with a particular trace, you can stop or even delete a trace configuration, preventing the tracing service from consuming unnecessary resources and disk space.
MCS provides two kinds of tracing: component and end-to-end. The trace style is specified when a new trace is configured, and can be updated later for an existing trace.
Component tracing is fairly straightforward; it applies to only one component of an MCS-managed product, for example, a runtime service. All activity for the specified component is traced and recorded for that component only, on a single computer or cluster.
End-to-end tracing, on the other hand, provides a method for tracing a particular task across several components, which may be installed on one or more computers or clusters.
MCS uses correlation IDs when it records trace files. For component-style traces, the correlation ID uniquely identifies each trace. For end-to-end trace files, the correlation ID is passed along with the task from one component to the next, and is used by MCS to correlate and display all of the related trace entries at the MCS console when requested. Each correlation ID also includes an index number, which is incremented by one at each stage of the task's progress. For component-style traces, the index number is always 1.
|Management & Control Services, Overview|
|Configuring Event Options|
|Creating Event Notifications|
|Configuring a New Trace|
|Starting or Stopping a Trace|