Parametrize prometheus rate duration
Depending on the ceilometer and prometheus configuration, it
may be required to change the duration used by the prometheus
rate function in autoscaling queries. The prefered value
for stable testing should be <ceilometer_polling_interval> +
<prometheus_scrape_interval>. With other values (especially with
values less than the ceilometer polling interval), it's possible,
that the rate() function will just return 0 for ever.
Change-Id: Ifbb6da5c5be9db6cba5494b6d28b24d28967ec1c
diff --git a/telemetry_tempest_plugin/config.py b/telemetry_tempest_plugin/config.py
index 8ff90aa..a478158 100644
--- a/telemetry_tempest_plugin/config.py
+++ b/telemetry_tempest_plugin/config.py
@@ -92,7 +92,11 @@
cfg.IntOpt('ceilometer_polling_interval',
default=300,
help="Polling interval configured for ceilometer. This can "
- "be used in test cases to wait for metrics to appear.")
+ "be used in test cases to wait for metrics to appear."),
+ cfg.IntOpt('prometheus_scrape_interval',
+ default=15,
+ help="Scrape interval configured for prometheus. This can "
+ "be used in test cases to properly configure autoscaling")
]
telemetry_services_opts = [