org.apache.beam.it.gcp

Class LoadTestBase

java.lang.Object

org.apache.beam.it.gcp.LoadTestBase

Direct Known Subclasses:

IOLoadTestBase

public abstract class LoadTestBase
extends java.lang.Object

Base class for performance tests. It provides helper methods for common operations.

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	LoadTestBase.MetricsConfiguration Utils for the metrics.

Field Summary

Fields

Modifier and Type	Field and Description
protected static com.google.auth.Credentials	CREDENTIALS
protected static com.google.api.gax.core.CredentialsProvider	CREDENTIALS_PROVIDER
protected MonitoringClient	monitoringClient
protected org.apache.beam.it.common.PipelineLauncher	pipelineLauncher
protected org.apache.beam.it.common.PipelineOperator	pipelineOperator
protected static java.lang.String	project
protected static java.lang.String	region
protected java.lang.String	testName
TestRule	watcher

Constructor Summary

Constructors

Constructor and Description
LoadTestBase()

Method Summary

Modifier and Type	Method and Description
static java.lang.Double	calculateAverage(java.util.List<java.lang.Double> values) Calculate the average from a series.
static org.apache.beam.it.common.PipelineOperator.Config	createConfig(org.apache.beam.it.common.PipelineLauncher.LaunchInfo info, java.time.Duration timeout)
protected void	exportMetricsToBigQuery(org.apache.beam.it.common.PipelineLauncher.LaunchInfo launchInfo, java.util.Map<java.lang.String,java.lang.Double> metrics) Exports the metrics of given dataflow job to BigQuery.
protected java.util.Map<java.lang.String,java.lang.Double>	getCpuUtilizationMetrics(java.lang.String jobId, com.google.monitoring.v3.TimeInterval timeInterval) Computes CPU Utilization metrics of the given job.
protected java.util.Map<java.lang.String,java.lang.Double>	getDataFreshnessMetrics(java.lang.String jobId, com.google.monitoring.v3.TimeInterval timeInterval) Computes Data freshness metrics of the given job.
protected java.util.Map<java.lang.String,java.lang.Double>	getMetrics(org.apache.beam.it.common.PipelineLauncher.LaunchInfo launchInfo) Computes the metrics of the given job using dataflow and monitoring clients.
protected java.util.Map<java.lang.String,java.lang.Double>	getMetrics(org.apache.beam.it.common.PipelineLauncher.LaunchInfo launchInfo, LoadTestBase.MetricsConfiguration config)
protected java.util.Map<java.lang.String,java.lang.Double>	getMetrics(org.apache.beam.it.common.PipelineLauncher.LaunchInfo launchInfo, java.lang.String inputPcollection) Computes the metrics of the given job using dataflow and monitoring clients.
protected java.util.Map<java.lang.String,java.lang.Double>	getMetrics(org.apache.beam.it.common.PipelineLauncher.LaunchInfo launchInfo, java.lang.String inputPcollection, java.lang.String outputPcollection) Computes the metrics of the given job using dataflow and monitoring clients.
protected java.util.Map<java.lang.String,java.lang.Double>	getSystemLatencyMetrics(java.lang.String jobId, com.google.monitoring.v3.TimeInterval timeInterval) Computes System latency metrics of the given job.
protected java.util.Map<java.lang.String,java.lang.Double>	getThroughputMetrics(org.apache.beam.it.common.PipelineLauncher.LaunchInfo jobInfo, LoadTestBase.MetricsConfiguration config, com.google.monitoring.v3.TimeInterval timeInterval) Computes throughput metrics of the given pcollection in dataflow job.
protected com.google.monitoring.v3.TimeInterval	getWorkerTimeInterval(org.apache.beam.it.common.PipelineLauncher.LaunchInfo info) Gets the time interval when workers were active to be used by monitoring client.
abstract org.apache.beam.it.common.PipelineLauncher	launcher()
void	setUp()
static void	setUpClass()
void	tearDownLoadTestBase()
protected boolean	waitForNumMessages(java.lang.String jobId, java.lang.String pcollection, java.lang.Long expectedElements) Checks if the input PCollection has the expected number of messages.

Methods inherited from class java.lang.Object

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

Field Detail

CREDENTIALS

protected static final com.google.auth.Credentials CREDENTIALS

CREDENTIALS_PROVIDER

protected static final com.google.api.gax.core.CredentialsProvider CREDENTIALS_PROVIDER

project

protected static java.lang.String project

region

protected static java.lang.String region

monitoringClient

protected MonitoringClient monitoringClient

pipelineLauncher

protected org.apache.beam.it.common.PipelineLauncher pipelineLauncher

pipelineOperator

protected org.apache.beam.it.common.PipelineOperator pipelineOperator

testName

protected java.lang.String testName

watcher

public TestRule watcher

Constructor Detail

LoadTestBase

public LoadTestBase()

Method Detail

setUpClass

public static void setUpClass()

setUp

public void setUp()
           throws java.io.IOException

Throws:

java.io.IOException

tearDownLoadTestBase

public void tearDownLoadTestBase()
                          throws java.io.IOException

Throws:

java.io.IOException

launcher

public abstract org.apache.beam.it.common.PipelineLauncher launcher()

exportMetricsToBigQuery

protected void exportMetricsToBigQuery(org.apache.beam.it.common.PipelineLauncher.LaunchInfo launchInfo,
                                       java.util.Map<java.lang.String,java.lang.Double> metrics)

Exports the metrics of given dataflow job to BigQuery.

Parameters:

launchInfo - Job info of the job

metrics - metrics to export

waitForNumMessages

protected boolean waitForNumMessages(java.lang.String jobId,
                                     java.lang.String pcollection,
                                     java.lang.Long expectedElements)

Checks if the input PCollection has the expected number of messages.

Parameters:

jobId - JobId of the job

pcollection - the input pcollection name

expectedElements - expected number of messages

Returns:

whether the input pcollection has the expected number of messages.

getMetrics

protected java.util.Map<java.lang.String,java.lang.Double> getMetrics(org.apache.beam.it.common.PipelineLauncher.LaunchInfo launchInfo,
                                                                      @Nullable
                                                                      java.lang.String inputPcollection,
                                                                      @Nullable
                                                                      java.lang.String outputPcollection)
                                                               throws java.lang.InterruptedException,
                                                                      java.io.IOException,
                                                                      java.text.ParseException

Computes the metrics of the given job using dataflow and monitoring clients.

Parameters:

launchInfo - Job info of the job

inputPcollection - input pcollection of the dataflow job to query additional metrics. If not provided, the metrics will not be calculated.

outputPcollection - output pcollection of the dataflow job to query additional metrics. If not provided, the metrics will not be calculated.

Returns:

metrics

Throws:

java.io.IOException - if there is an issue sending the request

java.text.ParseException - if timestamp is inaccurate

java.lang.InterruptedException - thrown if thread is interrupted

getMetrics

protected java.util.Map<java.lang.String,java.lang.Double> getMetrics(org.apache.beam.it.common.PipelineLauncher.LaunchInfo launchInfo,
                                                                      java.lang.String inputPcollection)
                                                               throws java.text.ParseException,
                                                                      java.lang.InterruptedException,
                                                                      java.io.IOException

Computes the metrics of the given job using dataflow and monitoring clients.

Parameters:

launchInfo - Job info of the job

inputPcollection - input pcollection of the dataflow job to query additional metrics. If not provided, the metrics will not be calculated.

Returns:

metrics

Throws:

java.io.IOException - if there is an issue sending the request

java.text.ParseException - if timestamp is inaccurate

java.lang.InterruptedException - thrown if thread is interrupted

getMetrics

protected java.util.Map<java.lang.String,java.lang.Double> getMetrics(org.apache.beam.it.common.PipelineLauncher.LaunchInfo launchInfo)
                                                               throws java.text.ParseException,
                                                                      java.lang.InterruptedException,
                                                                      java.io.IOException

Computes the metrics of the given job using dataflow and monitoring clients.

Parameters:

launchInfo - Job info of the job

Returns:

metrics

Throws:

java.io.IOException - if there is an issue sending the request

java.text.ParseException - if timestamp is inaccurate

java.lang.InterruptedException - thrown if thread is interrupted

getMetrics

protected java.util.Map<java.lang.String,java.lang.Double> getMetrics(org.apache.beam.it.common.PipelineLauncher.LaunchInfo launchInfo,
                                                                      LoadTestBase.MetricsConfiguration config)
                                                               throws java.io.IOException,
                                                                      java.lang.InterruptedException,
                                                                      java.text.ParseException

Throws:

java.io.IOException

java.lang.InterruptedException

java.text.ParseException

getCpuUtilizationMetrics

protected java.util.Map<java.lang.String,java.lang.Double> getCpuUtilizationMetrics(java.lang.String jobId,
                                                                                    com.google.monitoring.v3.TimeInterval timeInterval)

Computes CPU Utilization metrics of the given job.

Parameters:

jobId - dataflow job id

timeInterval - interval for the monitoring query

Returns:

CPU Utilization metrics of the job

getThroughputMetrics

protected java.util.Map<java.lang.String,java.lang.Double> getThroughputMetrics(org.apache.beam.it.common.PipelineLauncher.LaunchInfo jobInfo,
                                                                                LoadTestBase.MetricsConfiguration config,
                                                                                com.google.monitoring.v3.TimeInterval timeInterval)

Computes throughput metrics of the given pcollection in dataflow job.

Parameters:

jobInfo - dataflow job LaunchInfo

config - the

timeInterval - interval for the monitoring query

Returns:

throughput metrics of the pcollection

getDataFreshnessMetrics

protected java.util.Map<java.lang.String,java.lang.Double> getDataFreshnessMetrics(java.lang.String jobId,
                                                                                   com.google.monitoring.v3.TimeInterval timeInterval)

Computes Data freshness metrics of the given job.

Parameters:

jobId - dataflow job id

timeInterval - interval for the monitoring query

Returns:

Data freshness metrics of the job

getSystemLatencyMetrics

protected java.util.Map<java.lang.String,java.lang.Double> getSystemLatencyMetrics(java.lang.String jobId,
                                                                                   com.google.monitoring.v3.TimeInterval timeInterval)

Computes System latency metrics of the given job.

Parameters:

jobId - dataflow job id

timeInterval - interval for the monitoring query

Returns:

System latency metrics of the job

getWorkerTimeInterval

protected com.google.monitoring.v3.TimeInterval getWorkerTimeInterval(org.apache.beam.it.common.PipelineLauncher.LaunchInfo info)
                                                               throws java.text.ParseException

Gets the time interval when workers were active to be used by monitoring client.

Throws:

java.text.ParseException

calculateAverage

public static java.lang.Double calculateAverage(java.util.List<java.lang.Double> values)

Calculate the average from a series.

Parameters:

values - the input series.

Returns:

the averaged result.

createConfig

public static org.apache.beam.it.common.PipelineOperator.Config createConfig(org.apache.beam.it.common.PipelineLauncher.LaunchInfo info,
                                                                             java.time.Duration timeout)