Perform a metric query
A metric query returns the collected sample data from the various monitored instances. Depending on client requirements, you can present returned data in different ways.
Common query parameters
instanceType
-
Types of instances to get data from. Possible values include the following:
-
ds
-
proxy
-
sync
-
metrics-server
-
instanceLocation
-
Location(s) of the instances from which data is collected.
instanceHostname
-
Names of the machines hosting the instances.
instanceVersion
-
Version(s) of the instances providing the data.
instance
-
(multi-valued) – ID(s) of the instances from which data is collected. The instance ID is the
cn
of the external server and the same name as listed by thestatus
command. startTime
-
Include samples on or after the specified time. The time is either an absolute time in ISO 8601 format (such as 2016-08-13T19:36:00Z) or a time relative to the
endTime
(such as -5m or -4h). By default, the start time is -5m. endTime
-
Include samples on or before this time. The end time is either an absolute time in ISO 8601 format or a time relative to now (such as -5m or -4h). The default end time is now. Offset time values are relative to the current system clock time on the PingDataMetrics server.
maxIntervals
-
The number of separate intervals, between the start and end times, returned. This is considered the “resolution” of the data over time. By default, the maximum number of intervals is 1, which means all samples collected between the start and end times will be aggregated into one result according to the statistic selected.
statistic
-
Retrieve and apply this statistic to the data. Default for count based metrics is count and average for other metric types. Possible values include the following:
-
count
-
average
-
minimum
-
maximum
-
histogram
-
dimension
-
Include only these dimension values. A colon separates the dimension name and values, which are separated by commas (for example, op-type:add,delete).
pivot
-
Pivot by these dimensions. A pivot keeps the data separated along different dimensional values. The value “instance” can be used to keep the data separate between different instances. For metrics that have the histogram statistic, the histogram pivot can also be used to keep the values of each histogram bucket separate.
tz
-
Specifies the timezone to be used when displaying dates. By default, it is GMT. The timezone is specified in Java Time Zone format. For example, "US/Central" is CST in the United States.
Sub-parameters for the count and average statistics
Both the count and average statistics of count type metrics can have a rate scale applied to occurrences over a period of time using the per sub-parameter. The valid rate scaling values are:
-
s
orsecond
-
m
orminute
-
h
orhour
Sub-parameters for the histogram statistic
The histogram statistic includes all buckets and keeps the raw value for each bucket. Graphs can be configured to show the percentage of all operations above a given threshold, such as 50 ms. These graphs are useful for looking at a small percentage of operations in a given category. If the value falls between histogram bucket boundaries, the buckets where it falls will be included in the data. Possible values include the following:
min
-
Includes in the calculation only the histogram data above the given threshold.
max
-
Provides an upper bound on the histogram value.
percent
-
Allows the histogram values to be reported as a percentage of the overall values. Instead of returning raw counts, the value is a fraction of the total. This percentage is calculated within a pivot.
If both |