HP-UX Workload Manager User's Guide

Advanced WLM usage: Using performance metrics
Supplying data to WLM
Appendix H 493
For wlmrcvdc conceptual information, see “Sending data with wlmsend
and wlmrcvdc: How it works” on page 509. For wlmrcvdc syntax
information, see “wlmrcvdc” on page 388.
Existing metrics
If you already maintain application-specific metrics or system metrics,
you can send that data to WLM in one of two ways:
•Using the wlmsend utility
• Using the WLM API
Both the wlmsend utility and the API are explained in the section “What
methods exist for sending data to WLM?” on page 493.
What methods exist for sending data to WLM?
The section focuses on how you want to get data to WLM, then
determines what data is available for the given type of transport.
There are several methods for sending data to WLM:
• Sending data from the command line
• Sending data from a shell script
• Sending data from a perl program
• Sending data via stdout
• Sending data from a collector written in C
• Sending ARM transaction data from a modified C program
• Sending ARM transaction data from a script with simulated
transactions
As your data collector gathers data, it must send the data to WLM. The
easiest way to send data is by using the wlmrcvdc utility. Alternatively,
you can send this data using the WLM API.
How you send the data is generally decided by how you collect the data.
If you collect data using shell scripts or perl programs, the wlmsend and
wlmrcvdc utilities provide the most convenient route for sending data.
However, when using C programs to collect data, the WLM API is
typically better.