Scores

Scores are named metrics evaluated at the end of the experiment. Each [[score]] requires exactly one of taskName or componentName.

Fields

Field	Type	Required	Description
name	string	Yes	Score identifier (see table below)
taskName	string	No	Name of the task to measure
componentName	string	No	Name of the component to measure
threshold	float	Yes	Pass value. Lower-is-better scores must be at or below; higher-is-better must be at or above.

[[score]]
name = "ERR_RATE"
taskName = "traffic"
threshold = 5.0

[[score]]
name = "SERVICE_CPU"
componentName = "app"
threshold = 70.0

Score names

Name	Applies to	Unit	Direction	Description
ERR_RATE	traffic task	%	Lower is better	Peak HTTP error rate during the load test
LATENCY_95	traffic task	ms	Lower is better	Peak 95th percentile request latency during the load test
FUNC_TEST	apitest task	%	Higher is better	Percentage of API test cases that passed
SERVICE_CPU	service component	%	Lower is better	Peak CPU usage across service pods
SERVICE_MEMORY	service component	MB	Lower is better	Peak memory usage across service pods
REDIS_CPU	redis component	%	Lower is better	Peak CPU usage across Redis pods
MYSQL_CPU	mysql component	%	Lower is better	Peak CPU usage across MySQL pods
POSTGRESQL_CPU	postgres component	%	Lower is better	Peak CPU usage across PostgreSQL pods
POSTGRESQL_DISK_IO	postgres component	%	Lower is better	Peak disk I/O utilization across PostgreSQL pods
MEMCACHED_CPU	memcached component	%	Lower is better	Peak CPU usage across Memcached pods
CASSANDRA_CPU	cassandra component	%	Lower is better	Peak CPU usage across Cassandra nodes
CASSANDRA_DISK_IO	cassandra component	%	Lower is better	Peak disk I/O utilization across Cassandra nodes