/**
* Copyright 2013 Netflix, Inc.
* <p/>
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
* <p/>
* http://www.apache.org/licenses/LICENSE-2.0
* <p/>
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package com.netflix.servo.stats;
import java.util.Arrays;
/**
* Configuration options for a {@link com.netflix.servo.monitor.StatsTimer}
* <p/>
* By default we publish count (number of times the timer was executed), totalTime, and
* 95.0, and 99.0 percentiles.
* <p/>
* The size for the buffer used to store samples is controlled using the sampleSize field,
* and the frequency
* at which stats are computed is controlled with the computeFrequencyMillis option.
* By default these are
* set to 100,000 entries in the buffer, and computation at 60,000 ms (1 minute) intervals.
*/
public final class StatsConfig {
private static final String CLASS_NAME = StatsConfig.class.getCanonicalName();
private static final String SIZE_PROP = CLASS_NAME + ".sampleSize";
private static final String FREQ_PROP = CLASS_NAME + ".computeFreqMillis";
/**
* Builder for StatsConfig. By default the configuration includes count,
* total and 95th and 99th percentiles.
*/
public static class Builder {
private boolean publishCount = true;
private boolean publishTotal = true;
private boolean publishMin = false;
private boolean publishMax = false;
private boolean publishMean = false;
private boolean publishVariance = false;
private boolean publishStdDev = false;
private int sampleSize = Integer.parseInt(System.getProperty(SIZE_PROP, "1000"));
private long frequencyMillis = Long.parseLong(System.getProperty(FREQ_PROP, "60000"));
private double[] percentiles = {95.0, 99.0};
/**
* Whether to publish count or not.
*/
public Builder withPublishCount(boolean publishCount) {
this.publishCount = publishCount;
return this;
}
/**
* Whether to publish total or not.
*/
public Builder withPublishTotal(boolean publishTotal) {
this.publishTotal = publishTotal;
return this;
}
/**
* Whether to publish min or not.
*/
public Builder withPublishMin(boolean publishMin) {
this.publishMin = publishMin;
return this;
}
/**
* Whether to publish max or not.
*/
public Builder withPublishMax(boolean publishMax) {
this.publishMax = publishMax;
return this;
}
/**
* Whether to publish an average statistic or not. Note that if you plan
* to aggregate the values reported (for example across a cluster of nodes) you probably do
* not want to publish the average per node, and instead want to compute it by publishing
* total and count.
*/
public Builder withPublishMean(boolean publishMean) {
this.publishMean = publishMean;
return this;
}
/**
* Whether to publish variance or not.
*/
public Builder withPublishVariance(boolean publishVariance) {
this.publishVariance = publishVariance;
return this;
}
/**
* Whether to publish standard deviation or not.
*/
public Builder withPublishStdDev(boolean publishStdDev) {
this.publishStdDev = publishStdDev;
return this;
}
/**
* Set the percentiles to compute.
*
* @param percentiles An array of doubles describing which percentiles to compute. For
* example {@code {95.0, 99.0}}
*/
public Builder withPercentiles(double[] percentiles) {
this.percentiles = Arrays.copyOf(percentiles, percentiles.length);
return this;
}
/**
* Set the sample size.
*/
public Builder withSampleSize(int size) {
this.sampleSize = size;
return this;
}
/**
* How often to compute the statistics. Usually this will be set to the main
* poller interval. (Default is 60s.)
*/
public Builder withComputeFrequencyMillis(long frequencyMillis) {
this.frequencyMillis = frequencyMillis;
return this;
}
/**
* Create a new StatsConfig object.
*/
public StatsConfig build() {
return new StatsConfig(this);
}
}
private final boolean publishCount;
private final boolean publishTotal;
private final boolean publishMin;
private final boolean publishMax;
private final boolean publishMean;
private final boolean publishVariance;
private final boolean publishStdDev;
private final double[] percentiles;
private final int sampleSize;
private final long frequencyMillis;
/**
* Creates a new configuration object for stats gathering.
*/
public StatsConfig(Builder builder) {
this.publishCount = builder.publishCount;
this.publishTotal = builder.publishTotal;
this.publishMin = builder.publishMin;
this.publishMax = builder.publishMax;
this.publishMean = builder.publishMean;
this.publishVariance = builder.publishVariance;
this.publishStdDev = builder.publishStdDev;
this.sampleSize = builder.sampleSize;
this.frequencyMillis = builder.frequencyMillis;
this.percentiles = Arrays.copyOf(builder.percentiles, builder.percentiles.length);
}
/**
* Whether we should publish a 'count' statistic.
*/
public boolean getPublishCount() {
return publishCount;
}
/**
* Whether we should publish a 'totalTime' statistic.
*/
public boolean getPublishTotal() {
return publishTotal;
}
/**
* Whether we should publish a 'min' statistic.
*/
public boolean getPublishMin() {
return publishMin;
}
/**
* Whether we should publish a 'max' statistic.
*/
public boolean getPublishMax() {
return publishMax;
}
/**
* Whether we should publish an 'avg' statistic.
*/
public boolean getPublishMean() {
return publishMean;
}
/**
* Whether we should publish a 'variance' statistic.
*/
public boolean getPublishVariance() {
return publishVariance;
}
/**
* Whether we should publish a 'stdDev' statistic.
*/
public boolean getPublishStdDev() {
return publishStdDev;
}
/**
* Get the size of the buffer that we should use.
*/
public int getSampleSize() {
return sampleSize;
}
/**
* Get the frequency at which we should update all stats.
*/
public long getFrequencyMillis() {
return frequencyMillis;
}
/**
* Get a copy of the array that holds which percentiles we should compute. The percentiles
* are in the interval (0.0, 100.0)
*/
public double[] getPercentiles() {
return Arrays.copyOf(percentiles, percentiles.length);
}
/**
* {@inheritDoc}
*/
@Override
public String toString() {
return "StatsConfig{"
+ "publishCount=" + publishCount
+ ", publishTotal=" + publishTotal
+ ", publishMin=" + publishMin
+ ", publishMax=" + publishMax
+ ", publishMean=" + publishMean
+ ", publishVariance=" + publishVariance
+ ", publishStdDev=" + publishStdDev
+ ", percentiles=" + Arrays.toString(percentiles)
+ ", sampleSize=" + sampleSize
+ ", frequencyMillis=" + frequencyMillis
+ '}';
}
/**
* {@inheritDoc}
*/
@Override
public boolean equals(Object o) {
if (this == o) {
return true;
}
if (!(o instanceof StatsConfig)) {
return false;
}
final StatsConfig that = (StatsConfig) o;
return frequencyMillis == that.frequencyMillis
&& publishCount == that.publishCount
&& publishMax == that.publishMax
&& publishMean == that.publishMean
&& publishMin == that.publishMin
&& publishStdDev == that.publishStdDev
&& publishTotal == that.publishTotal
&& publishVariance == that.publishVariance
&& sampleSize == that.sampleSize
&& Arrays.equals(percentiles, that.percentiles);
}
/**
* {@inheritDoc}
*/
@Override
public int hashCode() {
int result = (publishCount ? 1 : 0);
result = 31 * result + (publishTotal ? 1 : 0);
result = 31 * result + (publishMin ? 1 : 0);
result = 31 * result + (publishMax ? 1 : 0);
result = 31 * result + (publishMean ? 1 : 0);
result = 31 * result + (publishVariance ? 1 : 0);
result = 31 * result + (publishStdDev ? 1 : 0);
result = 31 * result + Arrays.hashCode(percentiles);
result = 31 * result + sampleSize;
result = 31 * result + (int) (frequencyMillis ^ (frequencyMillis >>> 32));
return result;
}
}