==============
Microservices
==============
What is Monolith Architecture
-----------------------------
-> If we develop all the functionalities in single project then it is
called as Monolith architecture based application
-> We will package our application as a jar/war to deploy into server
-> As monolith application contains all functionalities, it will become
fat jar/war
Advantages
----------
1) Simple to develop
2) Everything is available at once place
3) Configuration required only once
Dis-Advantages
--------------
1) Difficult to maintain
2) Dependencies among the functionalites
3) Single Point Of Failure
4) Entire Project Re-Deployment
****** To overcome the problems of Monolithic, Microservices architecture
came into market******
-> Microservices is not a programming language
-> Microservices is not a framework
-> Microservices is not an API
###### Microservices is an architectural design pattern #####
-> Microservices suggesting to develop application functionalities with
loosely coupling
-> In Microservices architecture we don't develop all the functionalities
in single project. We will divide project functionalities into several
REST APIs.
*********************Note: One REST API is called as one
Microservice************************
-> Microservices architecture based project means collection of REST
APIs.
-> Microservices is not related to only java. Any programming language
specific project can use Microservices Architecture.
Advantages
----------
1) Loosely Coupling
2) Easy To maintain
3) Faster Development
4) Quick Deployment
5) Faster Releases
6) Less Downtime
7) Technology Independence (We can develop backend apis with multiple
technologies)
Dis-Advantages
--------------
1) Bounded Context (Deciding no.of services to be created)
2) Lot of configurations
3) Visibility
4) Pack of cards
===========================
Microservices Architecture
===========================
-> We don't have any fixed architecture for Microservices
-> People are customizing microservices architecture according to their
Project requirement
-> Most of the projects will use below components in Microservices
Architecture
1) Service Registry (Eureka Server)
2) Services (REST APIs)
3) Interservice Communication (FeginClient)
4) API Gateway
5) Admin Server
6) Zipkin
Service Registry
++++++++++++++++
-> Service Registry acts as DB of services available in the project
-> It provides the details of all the services which are registered with
Service Registry
-> We can identify how many services available in the project
-> We can identify how many instances available for each service
-> We can use "Eureka Server" as service registry
-> Eureka Server provided by "Spring Cloud Netflix" library
Services
+++++++++
-> Services means REST APIs / Microservices
-> Services contains backend business logic
-> In the project, some services will interact with DB
-> In the project, some services will interact with third party REST API
( external communication )
-> In the project, some services will interact with another services with
in the project
( inter-service communication )
-> For inter-service communication we will use feign-client
-> To distribute the load, we can run one service with Multiple Instances
(Load Balancing)
Note: We will register every service with Service Registry
API Gateway
+++++++++++
-> API Gateway is used to manage our backend apis of the project
-> API Gateway acts as mediator between end users and backend apis
-> API Gateway can contain filter logic to decide request processing
(Authentication)
-> API Gateway will contain Routing logic (which request should go to
which REST API)
-> API Gateway also will be registered with Service Registry
-> Spring Cloud Gateway we can use as API Gateway
Admin Server
------------
-> Admin Server is used to manage all backend apis actuator endpoints at
one place
-> Our backend apis will be registered with Admin Server
-> Admin Server will provide User interface to monitor apis actuator
endpoints
Zipin Server
---------------
-> Zipkin Server is used for Distributed Tracing
-> Using this Zipkin, we can monitor which API is taking more time to
process our request.
=============================================================
Mini Project Implementation using Microservices Architecture
=============================================================
STOCK-PRICE-API : It will maintain companies stock price details in db
table
-> Input : Company Name
-> Output : Company Stock Price Value
Sample Data
------------
TCS-3000
HCL-1500
AXIS-1200
STOCK-PRICE-CALC-API : It is used to calculate total stocks cost based on
company name and quantity
-> Input : Company Name & Quantity
-> Output : Total Stocks Cost
Note: To calculate total stocks cost, STOCK-PRICE-CALC-API should get
company stock price from STOCK-PRICE-API (INTER SERVICE COMMUNICATION)
===============================================================
Steps to develop Service Registry Application (Eureka Server)
===============================================================
1) Create Service Registry application with below dependency
- EurekaServer (spring-cloud-starter-netflix-eureka-server)
2) Configure @EnableEurekaServer annotation in boot start class
3) Configure below properties in application.yml file
server:
port: 8761
eureka:
client:
register-with-eureka: false
Note: If Service-Registry project port is 8761 then clients can discover
service-registry and will register automatically with service-registry.
If service-registry project running on any other port number then we have
to register clients with service-registry manually.
4) Once application started we can access Eureka Dashboard using below
URL
URL : http://localhost:8761/
Steps to develop stock-price-api (Eureka-Client-1)
++++++++++++++++++++++++++++++++++++++++++++++++++
1) Create Spring Boot application with below dependencies
- eureka-discovery-client
- starter-web
- starter-datajpa
- h2
- devtools
2) Configure @EnableDiscoveryClient annotation at start class
3) Configure below properties in application.yml file
server-port
h2-datasource-properties
application-name
-----------------------------------application.yml-----------------------
------------------
server:
port: 1111
spring:
application:
name: STOCK-PRICE-API
datasource:
username: sa
password: sa
url: jdbc:h2:mem:testdb
driver-class-name: org.h2.Driver
jpa:
defer-datasource-initialization: true
-------------------------------------------------------------------------
-------------------------
4) Create Entity class & Repository interface for COMPANY STOCK PRICE
DETAILS
5) Create "data.sql" file under src/main/resource folder with insert
queries to load data into db table like below
INSERT INTO STOCK_PRICE_DTLS (STOCK_ID, COMPANY_NAME, COMPANY_PRICE)
VALUES (101, 'TCS', 3000.00);
INSERT INTO STOCK_PRICE_DTLS (STOCK_ID, COMPANY_NAME, COMPANY_PRICE)
VALUES (102, 'HCL', 1500.00);
INSERT INTO STOCK_PRICE_DTLS (STOCK_ID, COMPANY_NAME, COMPANY_PRICE)
VALUES (103, 'HDFC', 4500.00);
INSERT INTO STOCK_PRICE_DTLS (STOCK_ID, COMPANY_NAME, COMPANY_PRICE)
VALUES (104, 'SBI', 450.00);
6) Create RestController to handle request & response
Input : CompanyName (Path Parameter)
Output : Stock Price Details
7) Run the application and check in Eureka Dashboard (It should display
in eureka dashboard)
Steps To Develop STOCK-PRICE-CALC-API (EUREKA-CLIENT-2)
+++++++++++++++++++++++++++++++++++++++++++++++++++++++
1) Create Spring Boot application with below dependencies
- web-starter
- devtools
- eureka-discovery-client
- fegin-client
2) Configure @EnableDiscoveryClient annotation at boot start class
3) Configure Server Port & application name in application.yml file
4) Create Rest Controller with required method
Input : Company Name & Quantity (Path Paramters)
Output : Total Cost
Note: In Rest Controller we should have logic to access STOCK-PRICE-API
-------------------------------------------------------------------------
--------------------------------
@RestController
public class StockCalcRestController {
@GetMapping("/calc/{cname}/{qty}")
public ResponseEntity<String> calculate(@PathVariable String cname,
@PathVariable Integer qty) {
String url = "http://localhost:1111/price/{cname}";
RestTemplate rt = new RestTemplate();
ResponseEntity<StockPrice> resEntity = rt.getForEntity(url,
StockPrice.class, cname);
StockPrice body = resEntity.getBody();
Double companyPrice = body.getCompanyPrice();
Double totalCost = companyPrice * qty;
String msg = "Total Cost : " + totalCost;
return new ResponseEntity<>(msg, HttpStatus.OK);
}
}
-------------------------------------------------------------------------
----------------------------------
-> In the above logic we have hard coded STOCK-PRICE-API url
-> If STOCK-PRICE-API url changed then calc-api logic should be changed
-> If STOCK-PRICE-API is running in Multiple instances for load balancing
our calc-api should access all the instances in round-robbin methodology
-> To overcome these problems we can use Interservice Communication using
FeignClient
-> Using FeginClient we can make rest call to another service using name
of the service (no need of url)
-> FeginClient will get service URL from service-registry based on
service-name
-------------------------------------------------------------------------
-------------------------------------
@FeignClient(name = "STOCK-PRICE-API")
public interface StockPriceClient {
@GetMapping("/price/{cname}")
public StockPrice invokeStockPrice(@PathVariable String cname);
}
-------------------------------------------------------------------------
------------------------------------
Note : Write @EnableFeignClients annotation at boot start class
-------------------------------------------------------------------------
------------------------------------
@RestController
public class StockCalcRestController {
@Autowired
private StockPriceClient priceClient;
@GetMapping("/calc/{cname}/{qty}")
public ResponseEntity<String> calculate(@PathVariable String cname,
@PathVariable Integer qty) {
StockPrice stockPrice = priceClient.invokeStockPrice(cname);
Double companyPrice = stockPrice.getCompanyPrice();
Double totalCost = companyPrice * qty;
String msg = "Total Cost : " + totalCost;
return new ResponseEntity<>(msg, HttpStatus.OK);
}
}
-------------------------------------------------------------------------
------------------------------------
Note: Run price-api with multiple instances using Run Configuration
Option
Note: Configure Port Number as VM Argument in Run Configuration
-Dserver.port=port-number
-------------------------------------------------------------------------
-------------------------------------
API Gateway
-------------------------------------------------------------------------
------------------------------------
-> API Gateway will act as mediator between client requests & backend
apis
-> API Gateway will provide single entrypoint to access our backend apis
-> In Api Gateway we will write mainley below 2 types of logics
1) Filters
2) Routing
-> Filters are used to execute some logic before request processing and
after request processing
-> Routing is used to tell which request should go to which REST API
-> In Spring Cloud we have 2 options to create API Gateway
1) Zuul Proxy (old approach)
2) Cloud Gateway (latest approach)
Note: Zuul Proxy is not supported by latest versions of spring boot
Working with Zuul proxy
++++++++++++++++++++++
-> create spring-boot application with below dependencies
1) Change Spring-boot-starter-parent version to 2.2.2.RELEASE
2) Change Cloud Version to "Hoxton.SR1"
3) Add zuul dependency
4) Add "eureka-client" dependency
-> Configure below 2 annotations in spring boot start class
@EnableZuulProxy
@EnableDiscoveryClient
application.yml file
+++++++++++++++++++++++
server:
port: 3333
spring:
application:
name: API-GATEWAY
zuul:
routes:
api1:
path: /stock/v1/**
service-id: STOCK-PRICE-API
api2:
path: /stock/v2/**
service-id: STOCK-CALC-API
-------------------------------------------------------------------------
-------------------------------------
-> We are running Service Registry project with Eureka Server on 8761
port number
-> Eureka Discovery Client applications are auto-registering with Eureka
Server when port is 8761
-> If we change Eureka Server port number then we have to register Eureka
Client application with Eureka Server using below property in
application.yml file
eureka:
client:
serviceUrl:
defaultZone: http://localhost:9090/eureka
Note: We should configure this property in eureka client application yml
file
-------------------------------------------------------------------------
-------------------------------------
Working with Spring Cloud API Gateway
-------------------------------------------------------------------------
------------------------------------
1) Create Spring boot application with below dependencies
-> web-stater
-> eureka-client
-> cloud-gateway
-> devtools
2) Configure @EnableDiscoveryClient annotation at boot start class
3) Configure API Gateway Routings in application.yml file like below
----------------------------------------application.yml file-------------
----------------------------------
spring:
cloud:
gateway:
discovery.locator:
enabled: true
lowerCaseServiceId: true
routes:
- id: stock-price-api
uri: lb://STOCK-PRICE-API
predicates:
- Path=/price/{companyName}
- id: stock-calc-api
uri: lb://STOCK-CALC-API
predicates:
- Path=/calc/{companyName}/{qty}
application:
name: CLOUD-API-GATEWAY
server:
port: 3333
-------------------------------------------------------------------------
-----------------------------------
In API gateway we will have 3 types of logics
1) Route
2) Predicate
3) Filters
-> Routing is used to defined which request should be processed by which
REST API in backend. Routes will be configured using Predicate
-> Predicate : This is a Java 8 Function Predicate. The input type is a
Spring Framework ServerWebExchange. This lets you match on anything from
the HTTP request, such as headers or parameters.
-> Filters are used to manipulate incoming request and outgoing response
of our application
Note: Using Filters we can implement security also for our application.
-------------------------------------------------------------------------
-------------------------------------
@Component
public class MyPreFilter implements GlobalFilter {
private Logger logger = LoggerFactory.getLogger(MyPreFilter.class);
@Override
public Mono<Void> filter(ServerWebExchange exchange,
GatewayFilterChain chain) {
logger.info("MyPreFilter :: filter () method executed...");
// Accessing HTTP Request information
ServerHttpRequest request = exchange.getRequest();
HttpHeaders headers = request.getHeaders();
Set<String> keySet = headers.keySet();
keySet.forEach(key -> {
List<String> values = headers.get(key);
System.out.println(key +" :: "+values);
});
return chain.filter(exchange);
}
}
-------------------------------------------------------------------------
---------------------------
@Component
public class MyPostFilter implements GlobalFilter {
final Logger logger = LoggerFactory.getLogger(MyPostFilter.class);
@Override
public Mono<Void> filter(ServerWebExchange exchange,
GatewayFilterChain chain) {
return chain.filter(exchange).then(Mono.fromRunnable(() -> {
logger.info("Global Post-filter executed...");
}));
}
}
-------------------------------------------------------------------------
---------------------------------
-> We can validate client given token in the request using Filter for
security purpose
-> We can write request and response tracking logic in Filter
-> Filters are used to manipulate request & response of our application
-> Any cross-cutting logics like security, logging, moniroing can be
implemented using Filters
===========================================
Spring Boot Admin Server and Admin Client
============================================
-> Admin server will provide user interface to monitor and manage all the
apis actuator endpoints
-> The REST APIs of our application should register with admin server (It
is called as Admin client)
Note: Using this approach we can monitor all the apis at one place
Working with Admin-Server
++++++++++++++++++++++++++
1) Create Boot application with admin-server dependency (select it while
creating the project)
2) Configure @EnableAdminServer annotation at start class
3) Run the boot application
4) Access application URL in browser (We can see Admin Server UI)
Working with Admin-Client
+++++++++++++++++++++++++
1) Create Spring Boot application with below dependecies
a) starter-web
b) starter-actuator
c) admin-starter-client
d) devtools
2) Configure below properties in application.yml file
a) server-port
b) application-name
c) enable-actuator-endpoints
d) configure admin serve URL to reiger
----------------------------------application.yml------------------------
-------------
server:
port: 1111
spring:
application:
name: CLIENT-ONE
boot:
admin:
client:
url: http://localhost:8080/
management:
endpoints:
web:
exposure:
include: '*'
-------------------------------------------------------------------------
------
3) Create Rest Controller with required methods
4) Run the appplication (It will register in Admin Server)
=================================
Working with Mono & Flux Objects
=================================
-> Mono & Flux objects are used to achieve reactive programming
-> Reacting Programming means Non-Blocking execution based on events
-> In Spring 5.x version Reactive Programming got introduced
-> To work with reactive programming Spring Boot provided below starter
'spring-boot-starter-webflux'
-> Mono object represents single response
-> Flux object represents stream of responses
-------------------------------Binding class-----------------------------
------------------
@Data
@NoArgsConstructor
@AllArgsConstructor
public class CustomerEvent {
private String name;
private Date eventDate;
}
-------------------------------RestController----------------------------
--------------------
@RestController
public class CustomerRestController {
@GetMapping("/event")
public Mono<CustomerEvent> getCustomerEvent() {
CustomerEvent event = new CustomerEvent("Smith", new Date());
Mono<CustomerEvent> mono = Mono.just(event);
return mono;
}
@GetMapping(value = "/events", produces =
MediaType.TEXT_EVENT_STREAM_VALUE)
public ResponseEntity<Flux<CustomerEvent>> getCustomerEvents() {
// Creating Customer data in the form of object
CustomerEvent event = new CustomerEvent("Smith", new Date());
// Create Stream object to send the data
Stream<CustomerEvent> customerStream = Stream.generate(() ->
event);
// Create Flux object with Stream
Flux<CustomerEvent> customerFlux =
Flux.fromStream(customerStream);
// Setting Response Interval
Flux<Long> interval = Flux.interval(Duration.ofSeconds(3));
// Combine Flux Interval and Customer Flux
Flux<Tuple2<Long, CustomerEvent>> zip = Flux.zip(interval,
customerFlux);
// Getting Flux value from the zip
Flux<CustomerEvent> fluxMap = zip.map(Tuple2::getT2);
// Returning Flux Response
return new ResponseEntity<>(fluxMap, HttpStatus.OK);
}
}
=================================
Circuit Breaker
=================================
-> Circuit Breaker is a design pattern in Microservices
-> Circuit Breaker is used to implement fault-tolerance systems
-> Fault-tolerance systems are also called as resillence systems
-> Fault-tolerance system means when main logic is failed to execute then
we should execute fallback logic to process client request
===========
Usecase
===========
=> Get data from redis, if redis logic is failing then we should get data
from database
Note: If redis logic is failing for 3 requests continuosly then execute
db logic for 30 mins. After 30 mins re-try for redis logic execution if
it is working then execute redis logic only. If 3 re-try executions
failed with redis then execute db logic for next 30 mins.
-> To implement circuit-breaker we should add below dependency in pom.xml
file
<dependency>
<groupId>io.pivotal.spring.cloud</groupId>
<artifactId>spring-cloud-services-starter-circuit-
breaker</artifactId>
</dependency>
-> Write @EnableHystrix annotation at boot start class
-------------------------------------------------------------------------
-----------------------
@RestController
public class DataRestController {
@GetMapping("/data")
@HystrixCommand(
fallbackMethod = "getDataFromDB",
commandProperties = {
@HystrixProperty(name="circuitBreaker.requestVolumeThreshold",
value="3"),
@HystrixProperty(name="circuitBreaker.sleepWindowInMilliseconds",
value="10000"),
@HystrixProperty(name="circuitBreaker.enabled", value="true")
}
)
public String getDataFromRedis() {
System.out.println("**getDataFromRedis() method called**");
if (new Random().nextInt(10) <= 10) {
throw new RuntimeException("Redis Server Is Down");
}
// logic to access data from redis
return "data accessed from redis (main logic) ....";
}
public String getDataFromDB() {
System.out.println("**getDataFromDB() method called**");
// logic to access data from db
return "data accessed from database (fall back logic) ....";
}
}
=================================
Sleuth & Zipkin
=================================
-> Microservices application means several REST APIs will be available
-> As part of application execution one Rest API can communicate another
REST API
-> When we send request from UI, it will process by Multiple REST APIs
with Interservice communication
*** How we can understand which rest api is taking more time to process
request ? ***
-> If we add Sleuth dependency in REST API then it will add span-id and
trace-id for log messages
-> For every request once span-id will be generated by Sleuth
-> If one request is processing multiple REST API then Sleuth will use
same span-id for REST APIs to generate log message
-> Trace-id is specific to one REST API
-> By using span-id and trace-id we can understand which REST api has
taken more time process request
-> To monitor span-id and trace-id details we will use ZipKin server
-> Zipkin server is providing user interface (UI) to monitor all the
details
Note: The REST APIs which are having sleuth dependency should register
with Zipkin server
Note: By using Sleuth and Zipkin we achieve Distributed Log Tracing
Steps to work with Sleuth and Zipkin
++++++++++++++++++++++++++++++++++++
1) create spring-boot application with below dependencies
a) web-starter
b) sleuth
c) zipkin
d) devtools
2) Create a REST Controller with required methods
3) Download zipkin-server jar file (https://zipkin.io/pages/quickstart)
4) Run zipkin-server using "java -jar <zipkin-jar-filename"
Note: Zipkin server runs on 9411 port
5) Run spring boot application and send a request to rest controller
method
6) Verify boot application logs display in console (span-id and trace-id
will be attached to logs)
7) Go to Zipkin server dashboard and monitor event details
( URL : http://localhost:9411 )
-------------------------------------------------------------------------
-------------------------
