HP-UX Workload Manager User's Guide

Example configuration files
stretch_goal.wlm
Chapter 9320
maxcpu = 50; # maximum CPU allocation (percentage)
entity = PRM group finance;
goal = metric fin_app.query.resp_time < 2.0;
condition = Mon - Fri; # only active on weekdays
}
# This is a “stretch” goal for the finance query group. If all other
# goals of higher priority (lower “pri” integral values) have been met,
# apply more CPU to group finance, so its application runs faster
# during prime time (Monday through Friday between 9am and 5pm).
#
slo finance_query_stretch {
pri = 5;
mincpu = 20;
maxcpu = 60; # let it take more CPU if available
entity = PRM group finance;
goal = metric fin_app.query.resp_time < 1.0;
condition = (Mon - Fri) && (09:00 - 17:00);
}
#
# This is an SLO for the Sales group, active at all times (there are no
# conditions or exceptions).
#
slo sales_query {
pri = 1;
mincpu = 40;
maxcpu = 80;
entity = PRM group sales;
goal = metric sales_app.resp_time < 10.0;
}
#
# tune structures
#
# This first structure specifies the data collector (coll_argv) for
# the fin_ap.query.resp_time metric, used in the goal statements
# of the finance SLOs (above). This hypothetical application
# (/opt/fin_app/finance_collector) is developed or otherwise provided
# by the user.
#
# NOTE: Because the data collectors for these metrics are
# hypothetical applications, this file will not pass the syntax-
# checking mode of wlmd (wlmd -c stretch_goal.wlm). A real data-
# collecting program must exist, as HP-UX WLM will launch it and rely
# upon it to provide performance metrics (or the HP-provided interface,
# wlmrcvdc, can be used when appropriate).