-
Sign up for a new account: https://newrelic.com/signup
-
Login and select ‘Get Started’
-
From ‘Choose your language’, select ‘Java’
-
Download the Java Agent zip file and unzip
-
Configure the license in
newrelic.yml
-
Set
enable_auto_app_naming
totrue
innewrelic.yml
. This will ensure that multiple microservices deployed in same instance of application server will be monitored in their own namespace. More details at: https://docs.newrelic.com/docs/agents/java-agent/configuration/automatic-application-naming
-
Download WildFly from http://download.jboss.org/wildfly/9.0.1.Final/wildfly-9.0.1.Final.zip.
-
WildFly installation for Java (standalone mode): https://docs.newrelic.com/docs/agents/java-agent/frameworks/wildfly-installation-java
-
Clone the workspace:
git clone https://github.com/arun-gupta/microservices.git
-
Start WildFly application server:
./bin/standalone.sh
-
Change the directory to
microservice
and deploy asmvn install -Pwildfly
-
Access the application at http://localhost:8080/everest-web
-
Go through the application flow and shop around
-
Access http://rpm.newrelic.com, select Applications to see the output
WarningThe data shown in the dashboard is updated after a few seconds as it is collected every minute. There is no way to change this setting for now. -
Select an application, say “everest-web”
-
Click on “Transactions”, “/Faces Servlet” to see more details about the application:
-
Scroll down and select on “Breakdown table” and show the % distribution
-
Click on the first row (as that is most time consuming resource). This is also the external microservice that is being invoked.
-
Click on “Service Maps” at the top to show how different services are interacting with each other:
WarningThis service map is a snapshot of how different microservices are interacting at the current moment. It does not provide historical data at this time.
TODO: . How to show the SQL query? . Response times . Errors and timeouts . Request flow . System resources (CPU, network, disk) . JVM pressure (threads GC, mem) . Apdex score