I need to read a different properties for Sentry.io in spring boot application based on environment profile (DEV, QA or PROD).
so using next properties for sentry in each bootstrap profile with different values
sentry:
enabled: true
environment: LOCAL
servername: localhost
dsn: MY_DNS_URL
stacktrace.app.packages:
- MY_PACKAGE
sample:
rate: 0.75
So how can i achieve that ??
Related
My Goal:
Having the Spring Cloud Config Server import active and provide a way for developers to have an optional property file on their machine.
My Config Server is up and running using org.springframework.cloud:spring-cloud-config-server:3.1.3 and #EnableConfigServer on the main class. Http requests to the concrete endpoint yield the expected result. This server should provide important environment configurations for a developer for his/her local setup.
$ curl http://localhost:8888/test-application/dev
{"name":"test-application","profiles":["dev"],"label":null,"version":null,"state":null,"propertySources":[{"name":"classpath:/cfg/test-application/application-dev.yml","source":{"server.port":1111}},{"name":"classpath:/cfg/test-application/application.yml","source":{"server.port":8000}}]}
Where localhost:8888 is my Config Server and test-application ist the name of the client application (defined via spring.application.name). The provided dev is the currently active profile on the client (The dev profile indicates a locally running software, there is no dev environment).
Clients configuration:
application.yml
spring:
application:
name: test-application
config:
import:
- configserver:http://localhost:8888 # <- pull the configuration from the configserver
- optional:file:/absolute/path/to/the/project/root/ # <- if there are any additional configuration files, use them
The client uses the following dependencies:
org.springframework.boot:spring-boot-starter-web:2.7.0
org.springframework.cloud:spring-cloud-starter-bootstrap:3.1.3
org.springframework.cloud:spring-cloud-starter:3.1.3
org.springframework.cloud:spring-cloud-starter-config:3.1.3
As shown above, the "base" application.yml configured server.port=8000 where the profile specific is set to server.port=1111. When reading the documentation this behaviour is correct. But my local developer configuration contains server.port=2222. This was ignored.
Here comes the problem:
When starting the client application, i can see the following log statements:
Fetching config from server at : http://localhost:8888
Located environment: name=test-application, profiles=[default], label=null, version=null, state=null
Located property source: [BootstrapPropertySource {name='bootstrapProperties-configClient'}, BootstrapPropertySource {name='bootstrapProperties-classpath:/cfg/test-application/application-dev.yml'}, BootstrapPropertySource {name='bootstrapProperties-classpath:/cfg/test-application/application.yml'}]
The following 1 profile is active: "dev"
Tomcat initialized with port(s): 1111 (http)
Initializing Spring embedded WebApplicationContext
Tomcat started on port(s): 1111 (http) with context path ''
Started TestApplicationKt in 2.234 seconds (JVM running for 2.762)
The configuration evaluation result from spring was to choose port 1111 instead of the wanted 2222 within the application-dev.yml located in the project root (developer config).
Wrapped up:
Three config files:
config server application.yml (port: 8000)
config server application-dev.yml (port: 1111)
project root developer config application-dev.yml (port: 2222) <- I want this file to have precedence over the other two.
When running the debugger, i see these found property sources within the injected Environment bean. The wanted developer file is within this list and the property source content is correct (server.port=2222).
Does anyone have an idea how to solve this problem?
I have created a project that reproduces this exact behaviour. Link to GitHub.
Thanks in advance!
Is the dev profile active when application is running in dev environment or it's also active when running locally?
Ideally you should have 2 different profiles for dev and local and you can enable config server only when dev profile is active.
You can rename your local property file as application-local.yml and disable spring cloud config when profile local is active.
You can put below code in bootstrap.yml(create new file beside application.yml) and remove config server configurations from application.yml
---
spring:
application:
name: test-application
config:
activate:
on-profile: dev
import:
- configserver:http://localhost:8888 # <- pull the configuration from the configserver
- optional:file:/absolute/path/to/the/project/root/ # <- if there are any additional configuration files, use them
---
spring:
application:
name: test-application
config:
activate:
on-profile: local
cloud:
config:
enabled: false
EDIT: Actual fix
Thanks for sharing reproducible example.
All you need to do is remove local application-dev.yml location reference from client application.yml and add it to configServer application.yml as shown in belolow screenshots.
/client/src/main/resources/application.yml
And put it in configServer application.yml at the end of search-locations. Order matters so please make sure you put overriding location/s at the end.
/configServer/src/main/resources/application.yml
Build and start both the applications and you will see client application server started on port 2222
I decided to deal with the profiles and divide into groups
spring:
//Common settings
---
spring:
profiles:
group:
default:
- prod
- actuator
dev:
- dev
- actuatorDev
uat:
- uat
- actuatorUat
Why when I specify dev or uat.
The settings are loaded from the main block, then replaced from dev or uat. And on top is the actuator profile
And when I don't specify the profile at startup, in theory it's just default
This kind of magic doesn't happen
How to correctly implement general settings and then replace them depending on the default, dev, uat profile?
Please Test:
//...
spring:
profiles:
default: "prod,actuator" # when no profile*S* set [3.]
group: # ...according to [3.1] and [3.2]
dev:
- "dev"
- "actuatorDev"
uat:
- "uat"
- "actuatorUat"
Spring (current) Ref: Chap. 3 Profiles
I am trying to read environment variable declared in deployment yaml of Kubernetes into springboot application.yaml
Below is sample in deployment.yaml
spec:
containers:
env:
- name: SECRET_IN
value: dev
Below is sample in application.yaml
innovation:
in: ${SECRET_IN:demo}
But on localhost when I try to print innovation.in (#Configuration is created correctly) I am not getting "dev" in output, it always prints demo, it appears the link between deployment, application yaml is not happening ,could someone please help.
You can store the whole application.YAML config file into the config map or secret and inject it with the deployment only
For example :
kind: ConfigMap
apiVersion: v1
metadata:
name: demo
data:
application.yaml: |-
pool:
size:
core: 1
max:16
if your application.properties is something like
example:
spring.datasource.url=jdbc:mysql://${MYSQL_HOST:localhost}:3306/dbname
spring.datasource.username=user
spring.datasource.password=password
You can replace it with
jdbc:mysql://${MYSQL_HOST:localhost}:3306/dbname
Deployment.yaml will be something like
spec:
containers:
- name: demowebapp
image: registry.gitlab.com/unicorn/unicornapp:1.0
ports:
- containerPort: 8080
imagePullPolicy: Always
env:
- name: MYSQL_HOST
value: mysql-prod
You can save more config into the config map & secret also based on the requirement.
Read more at : https://pushbuildtestdeploy.com/spring-boot-application.properties-in-kubernetes/
I think you did everything right, I have a similar working setup, although without a default 'demo'.
A couple of clarification from the spring boot's standpoint that might help.
application.yml can contain placeholders that can be resolved from the environment variables indeed.
Make sure that this application.yml is not "changed" (rewritten, filtered by maven whatever) during the compilation of the spring boot application artifact.
The most important: spring boot knows nothing about the k8s setup. If the environment variable exists - it will pick it. So the same could be checked even locally - define the env. variable on your local machine and run the spring boot application.
The chances are that somehow when the application runs (with the user/group) the environment variables are not accessible - check it by printing the environment variables (or this specific one) right before starting the spring boot application. Or you can do it in java in the main method:
Map<String, String> env = System.getenv();
env.entrySet().forEach(System.out::println);
I have facing an issue with Spring Cloud Config Server and Eureka Server Profiling.
Let's say I have 3 services with their name ("spring.application.name") as :
myapp-svc
myapp-spring-cloud-config-svc
myapp-spring-eureka-svc
I want to deploy each service in 2 regions ( dev and prod ). In Dev region, each service will run on localhost and in prod it will have some different url. 'myapp-spring-cloud-config-svc' in dev region will point to local git repo, while in prod region it will point to remote git repo.I can have 2 configurations:
1) When I start 'myapp-svc' service in local, it should connect to 'myapp-spring-cloud-config-svc' in dev.
I can do this by setting spring.cloud.config.uri = .
But the issue with this set up is that the property needs to be defined in bootstrap.properties.
So, If deploy 'myapp-svc' to prod, I will have to change config uri there to point it to prod config service which in turn would need another build creation.
This doesn't seem like a good solution, what if I have 50 app related services, I can't change this property in each one of them before prod deployment.
I tried setting spring.cloud.config.uri in application-dev.properties of 'myapp-svc' but it doesn't work. As per docs, it must be changed in bootstrap.
So, how do I implement this without having to create new build for prod ?
2) I can first call eureka and then using eureka I can call config service here.
The problem here is also same.
If I use eureka to look up config then "eureka.client.serviceUrl.defaultZone" must be defined in "bootstrap.yml".
See this:https://cloud.spring.io/spring-cloud-config/multi/multi__spring_cloud_config_client.html
So, in this case too, I need to change eureka url before deploying this service to prod.
Please help me on this...!!
Here is how, the properties, yml looks like for each of the above mentioned services:
1) myapp-svc:
1.1)bootstrap.yml
spring:
application:
name: myapp-svc
cloud:
config:
discovery:
enabled: true
serviceId: myapp-spring-cloud-config-svc
eureka:
client:
serviceUrl:
defaultZone: http://localhost:8762/eureka/
server:
port: 8082
2) myapp-spring-cloud-config-svc:
2.1)application-dev.properties:
spring.cloud.config.server.git.uri=file:///C:/config-repo
eureka.client.serviceUrl.defaultZone=http://localhost:8762/eureka
2.2)application-prod.properties:
spring.cloud.config.server.git.uri=https://github.com/<mygit Repo>
2.3)bootstrap.proerties:
spring.application.name=myapp-spring-cloud-config-svc
server.port=8888
3) myapp-spring-eureka-svc
3.1)bootstrap.proerties
spring.application.name=myapp-spring-eureka-svc
server.port=8762
1) You can have profile specific bootstrap-<profile>.properties (like for application-<profile>.properties) for each supported profile to avoid rebuilding your application for each env. Then just pass application profile using to your application during start-up. Spring will load correct bootstrap-<profile>.properties and will connect to proper configuration server (or eureka, etc). Example:
java -jar your-app.jar --spring.profiles.active=dev
2) You can pass your URLs externally as custom properties (same as with profile above) and have smth like this in bootstrap.properties. Example:
spring.cloud.config.uri=${config.server.url}
then pass --config.server.url= ... during start-up.
3) You can pass Spring properties in the same way during start-up. Example:
java -jar your-app.jar --spring.cloud.config.uri= ...
4) You can use system env variables. Example:
spring.cloud.config.uri=${SYSTEM_ENV_CLOUD_CONFIG_URI}
I am trying to cf push this code from this github repo after running mvn package, but I get a failed app deployment.
I have created a user defined service for mysql since cleardb is not available on my bosh-lite local vm. I created the mysql service and named it as 'mysql' as well as guided by this.
The application.yml in this spring boot app points to localhost. That doesn't work even for a local (non cf) build without corrrecting the jdbc url. Here is the stack trace from the bosh-lite vm when I try to cf push this app and read the logs using cf logs springpong --recent.
Also, why do we need both, the manifest.yml and the application.yml providing database details ?
vagrant#vagrant-ubuntu-trusty-64:~$ cat pong_matcher_spring/src/main/resources/application.yml
---
spring:
datasource:
driverClassName: com.mysql.jdbc.Driver
url: jdbc:mysql://localhost/pong_matcher_spring_development
username: springpong
password: springpong
jpa:
hibernate.ddl-auto: none
show_sql: false
vagrant#vagrant-ubuntu-trusty-64:~$ cat pong_matcher_spring/manifest.yml
---
applications:
- name: springpong
buildpack: java_buildpack
path: target/pong-matcher-spring-1.0.0.BUILD-SNAPSHOT.jar
services:
- mysql
vagrant#vagrant-ubuntu-trusty-64:~$
Can someone help me fix this ?
TIA
This was happening as the security groups were not setup to allow for network access. A sample gist of commands can be seen here -> https://gist.github.com/menicosia/2e9c414430138064f945
Caveat: This is only for bosh-lite and development purposes.