The microservice application performance problem analysis tool supports to locate performance bottlenecks via proposed intra/inter-trace analysis technique, and it can also display the analysis results visually.
1.Java-1.8.0
2.Python-3.8: Test execution environment
3.Nginx: front-end operating environment
4.Docker: container service
1.Hadoop-2.8.5, Hbase-1.4.13: Pinpoint storage service
2.Pinpoint-collector: Pinpoint collector
3.Pinpoint-web: Pinpoint interface
4.Pinpoint-agent: Pinpoint probe
msProf: Performance Analysis Service Project
1.Webflash-vue: front-end interface
2.Webflash: the background of the front end
3.Mysql-8.0: database
Trainticket: Application under test
1.Python
pip install locust==2.10.1
2.Nginx
sudo apt install nginx
1.Hadoop
1)Download Hadoop-2.8.5 installation package.
2)Modify core-site.xml in hadoop.
<configuration>
<property>
<name>fs.defaultFS</name>
<value>hdfs://localhost:9000</value>
</property>
<property>
<name>hadoop.tmp.dir</name>
<value>/root/softwares/hadoop/hadoop-2.7.1/hadoop_tmp</value>
</property>
</configuration>
3)Modify yarn-site.xml in hadoop
<configuration>
<!-- Site specific YARN configuration properties -->
<property>
<name>yarn.nodemanager.aux-services</name>
<value>mapreduce_shuffle</value>
</property>
<property>
<name>yarn.resourcemanager.hostname</name>
<value>localhost</value>
</property>
<property>
<name>yarn.nodemanager.resource.memory-mb</name>
<value>1024</value>
</property>
<property>
<name>yarn.nodemanager.resource.cpu-vcores</name>
<value>1</value>
</property>
</configuration>
4)Modify mapred-site.xml in hadoop
<configuration>
<property>
<name>mapreduce.framework.name</name>
<value>yarn</value>
</property>
</configuration>
5)Execute the command to initialize the file system
hdfs namenode -format
6)Add environment variables
export HADOOP_HOME="/root/softwares/hadoop/hadoop-2.7.1"
2.Hbase
1)Download Hbase-1.4.13 installation package
2)Modify hbase-site.xml in hbase
<configuration>
<property>
<name>hbase.rootdir</name>
<value>hdfs://localhost:9000/hbase</value>
</property>
<property>
<name>hbase.master.info.port</name>
<value>60010</value>
</property>
<property>
<name>hbase.zookeeper.property.dataDir</name>
<value>/root/softwares/hbase/hbase-1.2.7/zookeeper_data</value>
</property>
</configuration>
3)start Hbase
hbase/bin/start-hbase.sh
4)Hbase initialization script
hbase shell hbase-create.hbase
3.Pinpoint-collector
1)Upload /resource/pinpoint-collector-boot-2.2.0.jar to the server.
2)Start the application using the deployment script start.sh.
nohup java
-Dpinpoint.zookeeper.address=localhost -DMsD.span.url=http://127.0.0.1:8347/collect/span
-DMsD.spanChunk.url=http://127.0.0.1:8347/collect/spanChunk
-jar pinpoint-collector-boot-2.2.2.jar >nohup.out 2>&1 &
4.Pinpoint-web
1)Upload /resource/pinpoint-web-boot-2.2.0.jar to the server.
2)Start the application using the deployment script start.sh.
nohup java -jar -Dpinpoint.zookeeper.address=localhost pinpoint-web-boot-2.2.2.jar >nohup.out 2>&1 &
5.Pinpoint-agent
1)Upload /resource/pinpoint-agent-2.2.0.tar.gz to the server and decompress it.
2)Start the application using the deployment script start.sh.
java -javaagent:pinpoint-bootstrap-2.2.2.jar
-Dpinpoint.agentId={The application ID you set}
-Dpinpoint.applicationName={The application name you set}
-Dprofiler.collector.ip=collectorIP
-Dprofiler.transport.grpc.collector.ip= collectorIP
-Dprofiler.sampling.rate=1
-jar {Application under test}
1.msProf
1)Upload /resource/msProf-1.0.jar to the server.
2)Configure/msProf/src/main/resources/application.yaml, set test suite address, locust script address, har file address.
3)Configure /resources/locustExec.py,Change host to the address of your application under test.
4)Start the application using the deployment script start.sh.
nohup java -jar msProf-1.0.jar 2>&1 &
1.webflash-vue
1)Configure /etc/nginx/nginx.conf.
2)Start the Nginx service
2.webflash
1)Upload /resource/flash-api.jar to the server.
2)Configure /resources/application.properties.
SVCDIAGNOSER_URL=http://127.0.0.1:8347
3)Configure /resources/application-dev.properties.
spring.datasource.url=jdbc:mysql://127.0.0.1:3306/webflash?useSSL=false&useUnicode=true&characterEncoding=UTF8&
4)Start the application using the deployment script start.sh.
nohup java -jar flash-api.jar 2>&1 &
3.MySQL8.0
1)Upload /resource/webflash-mysql-8.0.tar to the server and decompress it.
2)Start and mount the directory with docker
docker run --name webflash-mysql-8.0 -e MYSQL_ROOT_PASSWORD=root -v /mysql/:/var/lib/mysql -p 33080:3306 -d webflash-mysql-8.0:pack
1.Trainticket
reference link:
1)Copy /resources/pinpoint-agent-2.2.2.tar.gz to each service under the /train-ticket-master/ directory.
2)Modify the start.sh in each service of Trainticket.
nohup java -Xmx200m
-javaagent:/pinpoint/pinpoint-agent-2.2.2/pinpoint-bootstrap-2.2.2.jar
-Dpinpoint.agentId={The application ID you set}
-Dpinpoint.applicationName={The application name you set}
-Dprofiler.collector.ip=127.0.0.1
-Dprofiler.transport.grpc.collector.ip=127.0.0.1
-Dprofiler.sampling.rate=1
-jar /app/ts-train-service-1.0.jar > /app/log/nohup.out 2>&1
3)compile project
mvn clean package -Dmaven.test.skip=true
4)build project
docker-compose build
5)start service
kubectl apply -f /root/train-ticket-2/k8s/ts-deployment-part1.yml
kubectl apply -f /root/train-ticket-2/k8s/ts-deployment-part2.yml
kubectl apply -f /root/train-ticket-2/k8s/ts-deployment-part3.yml
msProf supports structure analysis, execution time analysis and traffic anomaly analysis.
