org.apache.hadoop.metrics2.lib

Class MutableStat

java.lang.Object

org.apache.hadoop.metrics2.lib.MutableMetric

org.apache.hadoop.metrics2.lib.MutableStat

Direct Known Subclasses:

MutableRate

@InterfaceAudience.Public
 @InterfaceStability.Evolving
public class MutableStat
extends MutableMetric

A mutable metric with stats. Useful for keeping throughput/latency stats.

Constructor Summary

Constructors

Constructor and Description
MutableStat(String name, String description, String sampleName, String valueName) Construct a snapshot stat metric with extended stat off by default
MutableStat(String name, String description, String sampleName, String valueName, boolean extended) Construct a sample statistics metric

Method Summary

Modifier and Type	Method and Description
void	add(long value) Add a snapshot to the metric
void	add(long numSamples, long sum) Add a number of samples and their sum to the running stat Note that although use of this method will preserve accurate mean values, large values for numSamples may result in inaccurate variance values due to the use of a single step of the Welford method for variance calculation.
org.apache.hadoop.metrics2.util.SampleStat	lastStat() Return a SampleStat object that supports calls like StdDev and Mean.
void	resetMinMax() Reset the all time min max of the metric
void	setExtended(boolean extended) Set whether to display the extended stats (stdev, min/max etc.) or not
void	snapshot(MetricsRecordBuilder builder, boolean all) Get a snapshot of the metric

Methods inherited from class org.apache.hadoop.metrics2.lib.MutableMetric

changed, clearChanged, setChanged, snapshot

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

MutableStat

public MutableStat(String name,
                   String description,
                   String sampleName,
                   String valueName,
                   boolean extended)

Construct a sample statistics metric

Parameters:

name - of the metric

description - of the metric

sampleName - of the metric (e.g. "Ops")

valueName - of the metric (e.g. "Time", "Latency")

extended - create extended stats (stdev, min/max etc.) by default.

MutableStat

public MutableStat(String name,
                   String description,
                   String sampleName,
                   String valueName)

Construct a snapshot stat metric with extended stat off by default

Parameters:

name - of the metric

description - of the metric

sampleName - of the metric (e.g. "Ops")

valueName - of the metric (e.g. "Time", "Latency")

Method Detail

setExtended

public void setExtended(boolean extended)

Set whether to display the extended stats (stdev, min/max etc.) or not

Parameters:

extended - enable/disable displaying extended stats

add

public void add(long numSamples,
                long sum)

Add a number of samples and their sum to the running stat Note that although use of this method will preserve accurate mean values, large values for numSamples may result in inaccurate variance values due to the use of a single step of the Welford method for variance calculation.

Parameters:

numSamples - number of samples

sum - of the samples

add

public void add(long value)

Add a snapshot to the metric

Parameters:

value - of the metric

snapshot

public void snapshot(MetricsRecordBuilder builder,
                     boolean all)

Description copied from class: MutableMetric

Get a snapshot of the metric

Specified by:

snapshot in class MutableMetric

Parameters:

builder - the metrics record builder

all - if true, snapshot unchanged metrics as well

lastStat

public org.apache.hadoop.metrics2.util.SampleStat lastStat()

Return a SampleStat object that supports calls like StdDev and Mean.

Returns:

SampleStat

resetMinMax

public void resetMinMax()

Reset the all time min max of the metric