Agent Global Settings
The information in this section is specific to global settings for agents.
Agent Heartbeat Tab
Configuration Requests
Description: This setting specifies how often the agent requests configuration updates from the Management Server. Configuration updates include settings and Management Pack changes that are applicable to the managed computer. This interval also determines when agents receive tasks to run.
Default: 1 minute.
Override: No.
Impact: A higher value will result in a slightly lower demand on the Management Server, but this means the rules and configuration settings are not updated as often, and are not synchronized as often.
Caution: None.
Recommendations: For typical conditions, use the default setting. During time periods when you are updating or tuning the Management Packs, you might want to shorten the request interval so that agents receive changes sooner. If you do not anticipate that configuration changes will occur on a regular basis, you can lengthen the interval; however, performance gains will be minimal. Since tasks are received at this interval as well, there will be a delay if this value is increased.
Heartbeat Interval
Description: How often the agent sends a heartbeat to the Management Server.
Default: 10 seconds.
Override: No.
Impact: A shorter interval increases demand on the managed computer, network, and Management Server. A longer interval might result in missed heartbeat alerts, depending on the interval in the Management Pack rule.
Caution: Setting the interval below 10 seconds is not recommended, except for troubleshooting. The setting must be shorter than the Management Servers scan interval (which can be configured on the Management Servers/Heartbeat Checking tab). You cannot set this interval to 0.
Recommendations: For typical conditions, use the default setting. However, if you are troubleshooting connectivity problems with an agent, you might want to make the interval shorter to get more feedback after changes. If your network rarely experiences problems, or the Management Server is servicing many agents, then you might want to increase the interval to lower the load on the Management Server.
Responses Tab
Simultaneous responses allowed
Description: The number of rule-based responses that can be run simultaneously on the agent. This setting is global, but can be overridden per agent.
Default: 5 responses.
Override: Yes.
Impact: Increasing the setting uses more resources. Be sure to monitor for failed responses, and to increase the setting value, as needed. The value depends on the hardware resources, the services and applications running on the managed computer, network bandwidth, and other factors.
Caution: Setting the value too high can consume resources on the computer, which can cause it to hang or restart. Do not set this value to 0 because there will be no available queue threads which will cause the queues to fill.
Recommendations: For typical conditions, use the default setting. If the managed computer, typically, runs with higher resource consumption, you might want to lower the number of simultaneous responses to lower resource usage. If the managed computer rarely uses more than 50% CPU, and the rules that apply to the managed computer require a higher than average number of responses to run, such as file transfer responses, then you might want to increase the number of simultaneous responses to take advantage of the remaining resources.
Temporary Storage Tab
Maximum Disk Space
Description: The size of the agent cache, in kilobytes (KB), on the disk drive. This cache is used to store operations data temporarily, while the data is being sent to the Management Server.
Default: 3,000 KB.
Override: Yes.
Impact: You might want to increase the value if any of the following are true:
There are lengthy Management Server outages.
If the agent needs to process large bursts of event data.
Caution: Decreasing this value by too much might cause WMI or performance counter data to be dropped. Increasing this value by too much, without cause, might consume needed disk space. If the file must be greater than 10 megabyte (MB), this might indicate a problem with agent communications, or it could indicate that the agent is collecting too much data.
Recommendations: For typical conditions, use the default setting.
Service Monitoring Tab
Enable Service Monitoring
Description: Determines whether the agent checks and reports changes in the status of Windows services. Service Availability reports use this information.
Default: ON.
Override: Yes.
Impact: If ON, the agent will collect and send information to the Management Server about Windows services on the managed computer. If OFF, this information will not be collected and sent. This setting does not significantly affect the managed computers performance, unless that information is sent too often.
Caution: Many rules in the various Management Packs use the events this generates. As a side effect, the state view is updated. If the setting is OFF, the state view will work, but in approximately 80% of the cases it will not.
Recommendations: Unless the specific agent is running under conditions where service status is either not needed or not relevant, use the default setting, ON.
Status check interval
Description: How often the agent will collect information about the status of Windows services on the managed computer.
Default: 20 seconds.
Override: Yes.
Impact: Lowering the value might cause adverse affects on the performance of the managed computer. Increasing the value by too much might render the feature useless because status would be irrelevant for effective monitoring.
Caution: Lowering this value to below 10 might lead to performance problems for both the managed computer and the MOM agent.
Recommendations: For typical conditions, use the default setting. If the managed computer has many services, and the Management Server is servicing many agents, you might want to increase the interval to lower resource overhead on the agent and Management Server. If you are troubleshooting services on a specific agent, you might want to override the global setting and decrease the interval to get service state data more frequently.
Status reporting interval
Description: How often the agent will send information about the status of Windows services on the managed computer to the Management Server.
Override: None.
Impact: This value is automatically set to six times the Status checking interval. The agent collects six sets of data, and sends the average of the values. The multiplier cannot be configured with this interface, or in the registry.
Buffering Tab
Send event and performance data after
Description: How long event and performance data is cached on the managed computer before being sent to the Management Server.
Default: 30,000 milliseconds.
Override: Yes.
Impact: Setting this value to 0 causes event and performance data to be sent constantly. Increasing this value means that data is held until the packet size is reached OR the queue is full
Caution: Setting the value at 0 might cause the agent to use excessive network bandwidth, and more importantly, increase latency. In addition, having event data set lower that alert data will have a negative affect on the Service Level Agreement (SLA) for alert delivery because event delivery will have a relatively higher priority.
Setting the value too high will increase latency.
Recommendations: For typical conditions, use the default setting. If the managed computer has adequate resources you might want to increase this value. Typically, higher settings can improve bandwidth usage but increase latency.
Send alert and any associated responses after
Description: How long alerts, and any associated responses, are cached before being sent to the Management Server.
Default: 1,000 milliseconds.
Override: Yes.
Impact: Setting this value to 0 causes the alerts, and associated responses, to be sent constantly. Increasing this value means that the agent cache must hold more data before sending it and, therefore, must be larger. If you increase this value, you might also need to increase the Maximum disk space setting on the Temporary Storage tab.
Caution: Setting the value to 0 might cause the agent to use excessive network bandwidth. Setting the value too high might cause the agent cache to fill up, which will affect performance.
Recommendations: For typical conditions, use the default setting.
Communications Tab
Packet size
Description: The size of the network packets that are used to transmit data to the Management Server.
Default: 50KB.
Override: Yes.
Recommendations: For typical conditions, use the default setting. Increasing this size can improve bandwidth usage but increase latency. Configuring this setting too low will have a negative impact on both bandwidth usage and latency, depending upon how much data is sent.
Maximum amount of data per second
Description: The maximum amount of operations data sent, per second, from the agent to the Management Server. This setting, in conjunction with the Packet size setting, determines how many packets are sent, per second, to the Management Server.
Default: 1000 KB.
Override: Yes.
Impact: Increasing the value causes the agent to send more operations data, per second. Lowering the value causes the agent to send less operations data, per second.
Caution: Increasing this value to higher than 100,000 (100 MB/sec) might lead to network bandwidth problems, depending on your network speed. Lowering the value to below 100 (100 KB/sec) might cause the agent cache to fill up, which can result in operations data not being sent to the Management Server in a timely manner.
Note
MOM 2005 supports up to 50 agent-managed computers on a 128 KB connection.
Recommendations: For typical conditions, use the default setting. When determining how to tune this setting, consider the following:
The number of agents that could send data, simultaneously, on the network segment.
The total amount of operations data that the agents might send.
The total amount of network bandwidth that you want the MOM agent communications to use.
Event Collection Tab
Collect event binary data
Description: Determines whether the binary event data, for all events that have binary data, is sent to the Management Server.
Range: ON/OFF.
Default: OFF.
Override: Yes.
Impact: If ON, the agent will collect and send the binary event data to the Management Server for all events that have binary data. If OFF, this information will not be collected and sent. This setting can impact the managed computers performance significantly, especially if many events with binary data are collected.
Recommendations: For typical conditions, use the default setting, OFF.