The CDT Submission tab displays information about the complexity of data that agents collect and submit for Canonical Data Transform (CDT) processing. Excessively large values or sudden increases in this data can result in performance problems.Figure 16. CDT Submission tab
• CDT Node Count: The number of internal data nodes produced by the CDT. Typically, there is one node for each topology object and its metrics, and one global (root) node in the tree.This value takes into consideration the count of the topology object submission events that have associated observations or metrics. Topology objects that do not have any metrics associated with them may go through the CDT without affecting this count. Additionally, one topology object may be submitted several times with different metric values, then counted several times.
• CDT Value Count: The number of metrics and/or observations produced by the CDT.
• Max Stack Depth: The depth of the data tree structure submitted by each agent instance. When a topology object in the data submission refers to another topology object, that increases the depth by one. When the second object refers to a third, the depth is increased again.
• Node Count: The number of nodes in the submission. It is a measure of how large the data submission is.
• Pre-processing time: The Agent Manager makes two passes over the data submitted by an agent. This is the time spent on the first pass, when the Agent Manager fills in references, validates the submission, calculates missing timestamps, and performs other pre-processing tasks.
• Processing time: The length of time it takes the Agent Manager to transfer the data submission to the CDT processing engine of the Management Server.
The Queue Utilization tab contains graphs that tell you how well the client queue and cache are performing. It displays the count of messages in the queue over time, its size, and the rate at which it discards messages. It also shows the rates of cache-related operations and memory consumption, and the percentage of time the client spends on cache-related operations.Figure 17. Queue Utilization tab
This contains tips for deploying the Foglight™ Agent Manager to a large number of UNIX® or Windows® hosts in your environment. It also discusses some of the deployment options you might want to consider.Deploying the Foglight Agent Manager is a multi-step process that includes selecting and deploying an Agent Manager cartridge, downloading the Agent Manager installer, and installing the Agent Manager. The steps in this process are described in on page .If you are planning to deploy the Agent Manager to multiple UNIX hosts, you can create a deployment script to simplify this process. An example of this type of script is outlined below, along with suggestions for how you could customize the script to suit your environment.
These parameters are used with the --silent command option when you run the Agent Manager installer non-interactively—for example, to install the Agent Manager from the command-line onto remote machines, as described in the examples in this.