Posts Tagged Spring Boot
Eventual Consistency: Decoupling Microservices with Spring AMQP and RabbitMQ
Posted by Gary A. Stafford in DevOps, Java Development, Software Development on May 15, 2017
Introduction
In a recent post, Decoupling Microservices using Message-based RPC IPC, with Spring, RabbitMQ, and AMPQ, we moved away from synchronous REST HTTP for inter-process communications (IPC) toward message-based IPC. Moving to asynchronous message-based communications allows us to decouple services from one another. It makes it easier to build, test, and release our individual services. In that post, we did not achieve fully asynchronous communications. Although, we did achieve a higher level of service decoupling using message-based Remote Procedure Call (RPC) IPC.
In this post, we will fully decouple our services using the distributed computing model of eventual consistency. More specifically, we will use a message-based, event-driven, loosely-coupled, eventually consistent architectural approach for communications between services.
What is eventual consistency? One of the best definitions of eventual consistency I have read was posted on microservices.io. To paraphrase, ‘using an event-driven, eventually consistent approach, each service publishes an event whenever it updates its data. Other services subscribe to events. When an event is received, a service updates its data.’
Example of Eventual Consistency
Imagine Service A, the Customer service, inserts a new customer record into its database. Based on that ‘customer created’ event, Service A publishes a message containing the new customer object, serialized to JSON, to a lightweight persistent message queue.
Service B, the new customer Onboarding service, a subscriber to that queue, consumes Service A’s message. Service B then executes that same CRUD operation, inserting the same new customer record into its database.
In the above example, it can be said that the customer records in Service B’s database are eventually consistent with the customer records in Service A’s database. Service A makes a change and publishes a message in response to the event. Service B consumes the message and makes the same change. Eventually (likely milliseconds) Service B’s customer records are consistent with Service A’s customer records.
Why Eventual Consistency?
So what does this apparent added complexity and duplication of data buy us? Consider the advantages. Service B, the Onboarding service, requires no knowledge of, or a dependency on, Service A, the Customer service. Still, Service B has a current record of all the customers that Service A maintains. Instead of making repeated and potentially costly REST HTTP call or RPC message-based call to or from Service A to Service B for new customers, Service B queries its database for a list of customers.
The value of eventual consistency increases factorially as you scale a distributed system. Imagine dozens of distinct microservices, many requiring data from other microservices. Further, imagine multiple instances of each of those services all running in parallel. Decoupling services from one another, through asynchronous forms of IPC, messaging, and event-driven eventual consistency greatly simplifies the software development lifecycle and operations.
Demonstration
In this post, we could use a few different architectural patterns to demonstrate message passing with RabbitMQ and Spring AMQP. They including Work Queues, Publish/Subscribe, Routing, or Topics. To keep things as simple as possible, we will have a single Producer, publish messages to a single durable and persistent message queue. We will have a single Subscriber, a Consumer, consume the messages from that queue. We focus on a single type of event message.
Sample Code
To demonstrate Spring AMQP-based messaging with RabbitMQ, we will use a reference set of three Spring Boot microservices. The Election Service, Candidate Service, and Voter Service are all backed by MongoDB. The services and MongoDB, along with RabbitMQ and Voter API Gateway, are all part of the Voter API.
The Voter API Gateway, based on HAProxy, serves as a common entry point to all three services, as well as serving as a reverse proxy and load balancer. The API Gateway provides round-robin load-balanced access to multiple instances of each service.
All the source code found this post’s example is available on GitHub, within a few different project repositories. The Voter Service repository contains the Voter service source code, along with the scripts and Docker Compose files required to deploy the project. The Election Service repository, Candidate Service repository, and Voter API Gateway repository are also available on GitHub. There is also a new AngularJS/Node.js Web Client, to demonstrate how to use the Voter API.
For this post, you only need to clone the Voter Service repository.
Deploying Voter API
All components, including the Spring Boot services, MongoDB, RabbitMQ, API Gateway, and the Web Client, are individually deployed using Docker. Each component is publicly available as a Docker Image, on Docker Hub. The Voter Service repository contains scripts to deploy the entire set of Dockerized components, locally. The repository also contains optional scripts to provision a Docker Swarm, using Docker’s newer swarm mode, and deploy the components. We will only deploy the services locally for this post.
To clone and deploy the components locally, including the Spring Boot services, MongoDB, RabbitMQ, and the API Gateway, execute the following commands. If this is your first time running the commands, it may take a few minutes for your system to download all the required Docker Images from Docker Hub.
If everything was deployed successfully, you should observe six running Docker containers, similar to the output, below.
Using Voter API
The Voter Service, Election Service, and Candidate Service GitHub repositories each contain README files, which detail all the API endpoints each service exposes, and how to call them.
In addition to casting votes for candidates, the Voter service can simulate election results. Calling the /simulation endpoint, and indicating the desired election, the Voter service will randomly generate a number of votes for each candidate in that election. This will save us the burden of casting votes for this demonstration. However, the Voter service has no knowledge of elections or candidates. The Voter service depends on the Candidate service to obtain a list of candidates.
The Candidate service manages electoral candidates, their political affiliation, and the election in which they are running. Like the Voter service, the Candidate service also has a /simulation endpoint. The service will create a list of candidates based on the 2012 and 2016 US Presidential Elections. The simulation capability of the service saves us the burden of inputting candidates for this demonstration.
The Election service manages elections, their polling dates, and the type of election (federal, state, or local). Like the other services, the Election service also has a /simulation endpoint, which will create a list of sample elections. The Election service will not be discussed in this post’s demonstration. We will examine communications between the Candidate and Voter services, only.
REST HTTP Endpoint
As you recall from our previous post, Decoupling Microservices using Message-based RPC IPC, with Spring, RabbitMQ, and AMPQ, the Voter service exposes multiple, almost identical endpoints. Each endpoint uses a different means of IPC to retrieve candidates and generates random votes.
Calling the /voter/simulation/election/{election} endpoint and providing a specific election, prompts the Voter service to request a list of candidates from the Candidate service, based on the election parameter you input. This request is done using synchronous REST HTTP. The Voter service uses the HTTP GET method to request the data from the Candidate service. The Voter service then waits for a response.
The Candidate service receives the HTTP request. The Candidate service responds to the Voter service with a list of candidates in JSON format. The Voter service receives the response payload containing the list of candidates. The Voter service then proceeds to generate a random number of votes for each candidate in the list. Finally, each new vote object (MongoDB document) is written back to the vote collection in the Voter service’s voters database.
Message-based RPC Endpoint
Similarly, calling the /voter/simulation/rpc/election/{election} endpoint and providing a specific election, prompts the Voter service to request the same list of candidates. However, this time, the Voter service (the client) produces a request message and places in RabbitMQ’s voter.rpc.requests queue. The Voter service then waits for a response. The Voter service has no direct dependency on the Candidate service; it only depends on a response to its request message. In this way, it is still a synchronous form of IPC, but the Voter service is now decoupled from the Candidate service.
The request message is consumed by the Candidate service (the server), who is listening to that queue. In response, the Candidate service produces a message containing the list of candidates serialized to JSON. The Candidate service (the server) sends a response back to the Voter service (the client) through RabbitMQ. This is done using the Direct reply-to feature of RabbitMQ or using a unique response queue, specified in the reply-to header of the request message, sent by the Voter Service.
The Voter service receives the message containing the list of candidates. The Voter service deserializes the JSON payload to candidate objects. The Voter service then proceeds to generate a random number of votes for each candidate in the list. Finally, identical to the previous example, each new vote object (MongoDB document) is written back to the vote collection in the Voter service’s voters database.
New Endpoint
Calling the new /voter/simulation/db/election/{election} endpoint and providing a specific election, prompts the Voter service to query its own MongoDB database for a list of candidates.
But wait, where did the candidates come from? The Voter service didn’t call the Candidate service? The answer is message-based eventual consistency. Whenever a new candidate is created, using a REST HTTP POST request to the Candidate service’s /candidate/candidates endpoint, a Spring Data Rest Repository Event Handler responds. Responding to the candidate created event, the event handler publishes a message, containing a serialized JSON representation of the new candidate object, to a durable and persistent RabbitMQ queue.
The Voter service is listening to that queue. The Voter service consumes messages off the queue, deserializes the candidate object, and saves it to its own voters database, to the candidate collection. For this example, we are saving the incoming candidate object as is, with no transformations. The candidate object model for both services is identical.
When /voter/simulation/db/election/{election} endpoint is called, the Voter service queries its voters database for a list of candidates. They Voter service then proceeds to generate a random number of votes for each candidate in the list. Finally, identical to the previous two examples, each new vote object (MongoDB document) is written back to the vote collection in the Voter service’s voters database.
Exploring the Code
We will not review the REST HTTP or RPC IPC code in this post. It was covered in detail, in the previous post. Instead, we will explore the new code required for eventual consistency.
Spring Dependencies
To use AMQP with RabbitMQ, we need to add a project dependency on org.springframework.boot.spring-boot-starter-amqp. Below is a snippet from the Candidate service’s build.gradle file, showing project dependencies. The Voter service’s dependencies are identical.
AMQP Configuration
Next, we need to add a small amount of RabbitMQ AMQP configuration to both services. We accomplish this by using Spring’s @Configuration annotation on our configuration classes. Below is the abridged configuration class for the Voter service.
And here, the abridged configuration class for the Candidate service.
Event Handler
With our dependencies and configuration in place, we will define the CandidateEventHandler class. This class is annotated with the Spring Data Rest @RepositoryEventHandler and Spring’s @Component. The @Component annotation ensures the event handler is registered.
The class contains the handleCandidateSave method, which is annotated with the Spring Data Rest @HandleAfterCreate. The event handler acts on the Candidate object, which is the first parameter in the method signature.
Responding to the candidate created event, the event handler publishes a message, containing a serialized JSON representation of the new candidate object, to the candidates.queue queue. This was the queue we configured earlier.
Consuming Messages
Next, we let’s switch to the Voter service’s CandidateListService class. Below is an abridged version of the class with two new methods. First, the getCandidateMessage method listens to the candidates.queue queue. This was the queue we configured earlier. The method is annotated with theSpring AMQP Rabbit @RabbitListener annotation.
The getCandidateMessage retrieves the new candidate object from the message, deserializes the message’s JSON payload, maps it to the candidate object model and saves it to the Voter service’s database.
The second method, getCandidatesQueueDb, retrieves the candidates from the Voter service’s database. The method makes use of the Spring Data MongoDB Aggregation package to return a list of candidates from MongoDB.
RabbitMQ Management Console
The easiest way to observe what is happening with the messages is using the RabbitMQ Management Console. To access the console, point your web browser to localhost, on port 15672. The default login credentials for the console are guest/guest. As you successfully produce and consume messages with RabbitMQ, you should see activity on the Overview tab.
Recall we said the queue, in this example, was durable. That means messages will survive the RabbitMQ broker stopping and starting. In the below view of the RabbitMQ Management Console, note the six messages persisted in memory. The Candidate service produced the messages in response to six new candidates being created. However, the Voter service was not running, and therefore, could not consume the messages. In addition, the RabbitMQ server was restarted, after receiving the candidate messages. The messages were persisted and still present in the queue after the successful reboot of RabbitMQ.
Once RabbitMQ and the Voter service instance were back online, the Voter service successfully consumed the six waiting messages from the queue.
Service Logs
In addition to using the RabbitMQ Management Console, we may obverse communications between the two services by looking at the Voter and Candidate service’s logs. I have grabbed a snippet of both service’s logs and added a few comments to show where different processes are being executed.
First the Candidate service logs. We observe a REST HTTP POST request containing a new candidate. We then observe the creation of the new candidate object in the Candidate service’s database, followed by the event handler publishing a message on the queue. Finally, we observe the response is returned in reply to the initial REST HTTP POST request.
Now the Voter service logs. At the exact same second as the message and the response sent by the Candidate service, the Voter service consumes the message off the queue. The Voter service then deserializes the new candidate object and inserts it into its database.
MongoDB
Using the mongo Shell, we can observe six new 2016 Presidential Election candidates in the Candidate service’s database.
Now, looking at the Voter service’s database, we should find the same six 2016 Presidential Election candidates. Note the Object IDs are the same between the two service’s document sets, as are the rest of the fields (first name, last name, political party, and election). However, the class field is different between the two service’s records.
Production Considerations
The post demonstrated a simple example of message-based, event-driven eventual consistency. In an actual Production environment, there are a few things that must be considered.
- We only addressed a ‘candidate created’ event. We would also have to code for other types of events, such as a ‘candidate deleted’ event and an ‘candidate updated’ event.
- If a candidate is added, deleted, then re-added, are the events published and consumed in the right order? What about with multiple instances of the Voter service running? Does this pattern guarantee event ordering?
- How should the Candidate service react on startup if RabbitMQ is not available
- What if RabbitMQ fails after the Candidate services have started?
- How should the Candidate service react if a new candidate record is added to the database, but a ‘candidate created’ event message cannot be published to RabbitMQ? The two actions are not wrapped in a single transaction.
- In all of the above scenarios, what response should be returned to the API end user?
Conclusion
In this post, using eventual consistency, we successfully decoupled our two microservices and achieved asynchronous inter-process communications. Adopting a message-based, event-driven, loosely-coupled architecture, wherever possible, in combination with REST HTTP when it makes sense, will improve the overall manageability and scalability of a microservices-based platform.
References
- RabbitMQ: Understanding Message Broker
- Building Microservices: Inter-Process Communication in a Microservices Architecture
- Spring: Understanding AMQP
- AMQP 0-9-1 Complete Reference Guide
- Handling Eventual Consistency in JVM Microservices with Event Sourcing (javaone 2016)
All opinions in this post are my own and not necessarily the views of my current employer or their clients.
Decoupling Microservices using Message-based RPC IPC, with Spring, RabbitMQ, and AMPQ
Posted by Gary A. Stafford in DevOps, Enterprise Software Development, Java Development, Software Development on May 8, 2017
Introduction
There has been a considerable growth in modern, highly scalable, distributed application platforms, built around fine-grained RESTful microservices. Microservices generally use lightweight protocols to communicate with each other, such as HTTP, TCP, UDP, WebSockets, MQTT, and AMQP. Microservices commonly communicate with each other directly using REST-based HTTP, or indirectly, using messaging brokers.
There are several well-known, production-tested messaging queues, such as Apache Kafka, Apache ActiveMQ, Amazon Simple Queue Service (SQS), and Pivotal’s RabbitMQ. According to Pivotal, of these messaging brokers, RabbitMQ is the most widely deployed open source message broker.
RabbitMQ supports multiple messaging protocols. RabbitMQ’s primary protocol, the Advanced Message Queuing Protocol (AMQP), is an open standard wire-level protocol and semantic framework for high-performance enterprise messaging. According to Spring, ‘AMQP has exchanges, routes, and queues. Messages are first published to exchanges. Routes define on which queue(s) to pipe the message. Consumers subscribing to that queue then receive a copy of the message.’
Pivotal’s Spring AMQP project applies core Spring concepts to the development of AMQP-based messaging solutions. The project’s libraries facilitate management of AMQP resources while promoting the use of dependency injection and declarative configuration. The project provides a ‘template’ (RabbitTemplate) as a high-level abstraction for sending and receiving messages.
In this post, we will explore how to start moving Spring Boot Java services away from using synchronous REST HTTP for inter-process communications (IPC), and toward message-based IPC. Moving from synchronous IPC to messaging queues and asynchronous IPC decouples services from one another, allowing us to more easily build, test, and release individual microservices.
Message-Based RPC IPC
Decoupling services using asynchronous IPC is considered optimal by many enterprise software architects when developing modern distributed platforms. However, sometimes it is not easy or possible to get away from synchronous communications. Rightly or wrongly, often times services are architected, such that one service needs to retrieve data from another service or services, in order to process its own requests. It can be said, that service has a direct dependency on the other services. Many would argue, services, especially RESTful microservices, should not be coupled in this way.
There are several ways to break direct service-to-service dependencies using asynchronous IPC. We might implement request/async response REST HTTP-based IPC. We could also use publish/subscribe or publish/async response messaging queue-based IPC. These are all described by NGINX, in their article, Building Microservices: Inter-Process Communication in a Microservices Architecture; a must-read for anyone working with microservices. We might also implement an architecture which supports eventual consistency, eliminating the need for one service to obtain data from another service.
So what if we cannot implement asynchronous methods to break direct service dependencies, but we want to move toward message-based IPC? One answer is message-based Remote Procedure Call (RPC) IPC. I realize the mention of RPC might send cold shivers down the spine of many seasoned architected. Traditional RPC has several challenges, many which have been overcome with more modern architectural patterns.
According to Wikipedia, ‘in distributed computing, a remote procedure call (RPC) is when a computer program causes a procedure (subroutine) to execute in another address space (commonly on another computer on a shared network), which is coded as if it were a normal (local) procedure call, without the programmer explicitly coding the details for the remote interaction.’
Although still a form of RPC and not asynchronous, it is possible to replace REST HTTP IPC with message-based RPC IPC. Using message-based RPC, services have no direct dependencies on other services. A service only depends on a response to a message request it makes to that queue. The services are now decoupled from one another. The requestor service (the client) has no direct knowledge of the respondent service (the server).
RPC with RabbitMQ and AMQP
RabbitMQ has an excellent set of six tutorials, which cover the basics of creating messaging applications, applying different architectural patterns, using RabbitMQ, in several different programming languages. The sixth and final tutorial covers using RabbitMQ for RPC-based IPC, with the request/reply architectural pattern.
Pivotal recently added Spring AMPQ implementations to each RabbitMQ tutorial, based on their Spring AMQP project. If you recall, the Spring AMQP project applies core Spring concepts to the development of AMQP-based messaging solutions.
This post’s RPC IPC example is closely based on the architectural pattern found in the Spring AMQP RabbitMQ tutorial.
Sample Code
To demonstrate Spring AMQP-based RPC IPC messaging with RabbitMQ, we will use a pair of simple Spring Boot microservices. These services, the Voter and Candidate services, have been used in several previous posts, and for training and testing DevOps engineers. Both services are backed by MongoDB. The services and MongoDB, along with RabbitMQ, are all part of the Voter API project. The Voter API project also contains an HAProxy-based API Gateway, which provides indirect, load-balanced access to the two services.
All code necessary to build this post’s example is available on GitHub, within three projects. The Voter Service project repository contains the Voter service source code, along with the scripts and Docker Compose files required to deploy the project. The Candidate Service project repository and the Voter API Gateway project repository are also available on GitHub. For this post, you need only clone the Voter Service project repository.
Deploying Voter API
All components, including the two Spring services, MongoDB, RabbitMQ, and the API Gateway, are individually deployed using Docker. Each component is publicly available as a Docker Image, on Docker Hub.
The Voter Service repository contains scripts to deploy the entire set of Dockerized components, locally. The repository also contains optional scripts to provision a Docker Swarm, using Docker’s newer swarm mode, and deploy the components. We will only deploy the services locally for this post.
To clone and deploy the components locally, including the two Spring services, MongoDB, RabbitMQ, and the API Gateway, execute the following commands. If this is your first time running the commands, it may take a few minutes for your system to download all the required Docker Images from Docker Hub.
If everything was deployed successfully, you should see the following output. You should observe five running Docker containers.
Using Voter API
The Voter Service and Candidate Service GitHub repositories both contain README files, which detail all the API endpoints each service exposes, and how to call them.
In addition to casting votes for candidates, the Voter service has the ability to simulate election results. By calling a /simulation endpoint, and indicating the desired election, the Voter service will randomly generate a number of votes for each candidate in that election. This will save us the burden of casting votes for this demonstration. However, the Voter service has no knowledge of elections or candidates. To obtain a list of candidates, the Voter service depends on the Candidate service.
The Candidate service manages electoral candidates, their political affiliation, and the election in which they are running. Like the Voter service, the Candidate service also has a /simulation endpoint. The service will create a list of candidates based on the 2012 and 2016 US Presidential Elections. The simulation capability of the service saves us the burden of inputting candidates for this demonstration.
REST HTTP Endpoint
The Voter service exposes two almost identical endpoints. Both endpoints generate random votes. However, below the covers, the two endpoints are very different. Calling the /voter/simulation/election/{election} endpoint, prompts the Voter service to request a list of candidates from the Candidate service, based on the election parameter you input. This request is done using synchronous REST HTTP. The Voter service uses the HTTP GET method to request the data from the Candidate service. The Voter service then waits for a response.
The HTTP request is received by the Candidate service. The Candidate service responds to the Voter service with a list of candidates, in JSON format. The Voter service receives the response containing the list of candidates. The Voter service then proceeds to generate a random number of votes for each candidate. Finally, each new vote object (MongoDB document) is written back to the vote collection in the Voter service’s voters database.
Message-based RPC Endpoint
Similarly, calling the /voter/simulation/rpc/election/{election} endpoint with a specific election prompts the Voter service to request the same list of candidates. However, this time, the Voter service (the client), produces a request message and places in RabbitMQ’s voter.rpc.requests queue. The Voter service then waits for a response. The Voter service has no direct dependency on the Candidate service. It only depends on a response to its message request. In this way, it is still a synchronous form of IPC, but the Voter service is now decoupled from the Candidate service.
The request message is consumed by the Candidate service (the server), who is listening to that queue. In response, the Candidate service produces a message containing the list of candidates, serialized to JSON. The Candidate service (the server) sends a response back to the Voter service (the client), through RabbitMQ. This is done using the Direct reply-to feature of RabbitMQ, or using a unique response queue, specified in the reply-to header of the request message, sent by the Voter Service.
According to RabbitMQ, ‘the direct reply-to feature allows RPC clients to receive replies directly from their RPC server, without going through a reply queue. (“Directly” here still means going through AMQP and the RabbitMQ server; there is no separate network connection between RPC client and RPC server.)’
According to Spring, ‘starting with version 3.4.0, the RabbitMQ server now supports Direct reply-to; this eliminates the main reason for a fixed reply queue (to avoid the need to create a temporary queue for each request). Starting with Spring AMQP version 1.4.1 Direct reply-to will be used by default (if supported by the server) instead of creating temporary reply queues. When no replyQueue is provided (or it is set with the name amq.rabbitmq.reply-to), the RabbitTemplate will automatically detect whether Direct reply-to is supported and either uses it or fall back to using a temporary reply queue. When using Direct reply-to, a reply-listener is not required and should not be configured.’ We are using the latest versions of both RabbitMQ and Spring AMQP, which should support Direct reply-to.
The Voter service receives the message containing the list of candidates. The Voter service deserializes the JSON payload to Candidate objects and proceeds to generate a random number of votes for each candidate in the list. Finally, each new vote object (MongoDB document) is written back to the vote collection in the Voter service’s voters database.
Exploring the RPC Code
We will not examine the REST HTTP IPC code in this post. Instead, we will explore the RPC code. You are welcome to download the source code and explore the REST HTTP code pattern; it uses some advanced features of Spring Boot and Spring Data.
Spring Dependencies
In order to use RabbitMQ, we need to add a project dependency on org.springframework.boot.spring-boot-starter-amqp. Below is a snippet from the Candidate service’s build.gradle file, showing project dependencies. The Voter service’s dependencies are identical.
AMQP Configuration
Next, we need to add a small amount of RabbitMQ AMQP configuration to both services. We accomplish this by using Spring’s @Configuration annotation on our configuration classes. Below is the configuration class for the Voter service.
And here, the configuration class for the Candidate service.
Candidate Service Code
With the dependencies and configuration in place, we define the method in the Voter service, which will request the candidates from the Candidate service, using RabbitMQ. Below is an abridged version of the Voter service’s CandidateListService class, containing the getCandidatesMessageRpc method. This method calls the rabbitTemplate.convertSendAndReceive method (see line 5, below).
Voter Service Code
Next, we define a method in the Candidate service, which will process the Voter service’s request. Below is an abridged version of the CandidateController class, containing the getCandidatesMessageRpc method. This method is decorated with Spring’s @RabbitListener annotation (see line 1, below). This annotation marks c to be the target of a Rabbit message listener on the voter.rpc.requests queue.
Also shown, are the getCandidatesMessageRpc method’s two helper methods, getByElection and serializeToJson. These methods query MongoDB for the list of candidates and serialize the list to JSON.
Demonstration
To demonstrate both the synchronous REST HTTP IPC code and the Spring AMQP-based RPC IPC code, we will make a few REST HTTP calls to the Voter API Gateway. For convenience, I have provided a shell script, demostrate_ipc.sh, which executes all the API calls necessary. I have added sleep commands to slow the output to the terminal down a bit, for easier analysis. The script requires HTTPie, a great time saver when working with RESTful services.
The demostrate_ipc.sh script does three things. First, it calls the Candidate service to generate a group of sample candidates. Next, the script calls the Voter service to simulate votes, using synchronous REST HTTP. Lastly, the script repeats the voter simulation, this time using message-based RPC IPC. All API calls are done through the Voter API Gateway on port 8080. To understand the API calls, examine the script, below.
Below is the list of candidates for the 2016 Presidential Election, generated by the Candidate service. The JSON payload was retrieved using the Voter service’s /voter/candidates/rpc/election/{election} endpoint. This endpoint uses the same RPC IPC method as the Voter service’s /voter/simulation/rpc/election/{election} endpoint.
Based on the list of candidates, below are the simulated election results. This JSON payload was retrieved using the Voter service’s /voter/results endpoint.
RabbitMQ Management Console
The easiest way to observe what is happening with our messages is using the RabbitMQ Management Console. To access the console, point your web-browser to localhost, on port 15672. The default login credentials for the console are guest/guest.
As you successfully send and receive messages between the services through RabbitMQ, you should see activity on the Overview tab. In addition, you should see a number of Connections, Channels, Exchanges, Queues, and Consumers.
In the Queues tab, you should find a single queue, the voter.rpc.requests queue. This queue was configured in the Candidate service’s configuration class, shown previously.
In the Exchanges tab, you should see one exchange, voter.rpc, which we configured in both the Voter and the Candidate service’s configuration classes (aka DirectExchange). Also, visible in the Exchanges tab, should be the routing key rpc, which we configured in the Candidate service’s configuration class (aka Binding).
The route binds the exchange to the voter.rpc.requests queue. If you recall Spring’s description, AMQP has exchanges (DirectExchange), routes (Binding), and queues (Queue). Messages are first published to exchanges. Routes define on which queue(s) to pipe the message. Consumers subscribing to that queue then receive a copy of the message.
In the Channels tab, you should note two connections, the single instances of the Voter and Candidate services. Likewise, there are two channels, one for each service. You can differentiate the channels by the presence of the consumer tag. The consumer tag, in this example, amq.ctag-Anv7GXs7ZWVoznO64euyjQ, uniquely identifies the consumer. In this example, the Voter service is the consumer. For a more complete explanation of the consumer tag, check out RabbitMQ’s AMQP documentation.
Message Structure
Messages cannot be viewed directly in the RabbitMQ Management Console. One way I have found to view messages is using your IDE’s debugger. Below, I have added a breakpoint on the Candidate service’ getCandidatesMessageRpc method, using IntelliJ IDEA. You can view the Voter service’s request message, as it is received by the Candidate service.
Note the message payload, the requested election. Note the twelve message header elements. The headers include the AMQP exchange, queue, and binding. The message headers also include the consumer tag. The message also uniquely identifies the reply-to queue to use, if the server does not support Direct reply-to (see earlier explanation).
Service Logs
In addition to the RabbitMQ Management Console, we may obverse communications between the two services, by looking at the Voter and Candidate service’s logs. I have grabbed a snippet of both service’s logs and added a few comments to show where different processes are being executed. First the Voter service logs.
Next, the Candidate service logs.
Performance
What about the performance of Spring AMQP RPC IPC versus REST HTTP IPC? RabbitMQ has proven to be very performant, having been clocked at one million messages per second on GCE. I performed a series of fairly ‘unscientific’ performance tests, completing 250, 500, and then 1,000 requests. The tests were performed on a six-node Docker Swarm cluster with three instances of each service in a round-robin load-balanced configuration, and a single instance of RabbitMQ. The scripts to create the swarm cluster can be found in the Voter service GitHub project.
Based on consistent test results, the speed of the two methods was almost identical. Both methods performed between 3.1 to 3.2 responses per second. For example, the Spring AMQP RPC IPC method successfully completed 1,000 requests in 5 minutes and 11 seconds, while the REST HTTP IPC method successfully completed 1,000 requests in 5 minutes and 18 seconds, 7 seconds slower than the RPC method.
There are many variables to consider, which could dramatically impact IPC performance. For example, RabbitMQ was not clustered. Also, we did not use any type of caching, such as Varnish, Memcached, or Redis. Both these could dramatically increase IPC performance.
There are also several notable differences between the two methods from a code perspective. The REST HTTP method relies on Spring Data Projection combined with Spring Data MongoDB Repository, to obtain the candidate list from MongoDB. Somewhat differently, the RPC method makes use of Spring Data MongoDB Aggregation to return a list of candidates. Therefore, the test results should be taken with a grain of salt.
Production Considerations
The post demonstrated a simple example of RPC communications between two services using Spring AMQP. In an actual Production environment, there are a few things that must be considered, as Pivotal points out:
- How should either service react on startup if RabbitMQ is not available? What if RabbitMQ fails after the services have started?
- How should the Voter server (the client) react if there are no Candidate service instances (the server) running?
- Should the Voter service have a timeout for the RPC response to return? What should happen if the request times out?
- If the Candidate service malfunctions and raises an exception, should it be forwarded to the Voter service?
- How does the Voter service protect against invalid incoming messages (eg checking bounds of the candidate list) before processing?
- In all of the above scenarios, what, if any, response is returned to the API end user?
Conclusion
Although in this post we did not achieve asynchronous inter-process communications, we did achieve a higher level of service decoupling, using message-based RPC IPC. Adopting a message-based, loosely-coupled architecture, whether asynchronous or synchronous, wherever possible, will improve the overall functionality and deliverability of a microservices-based platform.
References
- RabbitMQ: Understanding Message Broker
- Remote Procedure Call (RPC) Using the Spring AMQP Client
- Building Microservices: Inter-Process Communication in a Microservices Architecture
- Spring: Understanding AMQP
- AMQP 0-9-1 Complete Reference Guide
All opinions in this post are my own and not necessarily the views of my current employer or their clients.
Streaming Docker Logs to Elastic Stack (ELK) using Fluentd
Posted by Gary A. Stafford in Bash Scripting, DevOps, Enterprise Software Development, Software Development on April 10, 2017
Introduction
Fluentd and Docker’s native logging driver for Fluentd makes it easy to stream Docker logs from multiple running containers to the Elastic Stack. In this post, we will use Fluentd to stream Docker logs from multiple instances of a Dockerized Spring Boot RESTful service and MongoDB, to the Elastic Stack (ELK).
In a recent post, Distributed Service Configuration with Consul, Spring Cloud, and Docker, we built a Consul cluster using Docker swarm mode, to host distributed configurations for a Spring Boot service. We will use the resulting swarm cluster from the previous post as a foundation for this post.
Fluentd
According to the Fluentd website, Fluentd is described as an open source data collector, which unifies data collection and consumption for a better use and understanding of data. Fluentd combines all facets of processing log data: collecting, filtering, buffering, and outputting logs across multiple sources and destinations. Fluentd structures data as JSON as much as possible.
Logging Drivers
Docker includes multiple logging mechanisms to get logs from running containers and services. These mechanisms are called logging drivers. Fluentd is one of the ten current Docker logging drivers. According to Docker, The fluentd logging driver sends container logs to the Fluentd collector as structured log data. Then, users can utilize any of the various output plugins, from Fluentd, to write these logs to various destinations.
Elastic Stack
The ELK Stack, now known as the Elastic Stack, is the combination of Elastic’s very popular products: Elasticsearch, Logstash, and Kibana. According to Elastic, the Elastic Stack provides real-time insights from almost any type of structured and unstructured data source.
Setup
All code for this post has been tested on both MacOS and Linux. For this post, I am provisioning and deploying to a Linux workstation, running the most recent release of Fedora and Oracle VirtualBox. If you want to use AWS or another infrastructure provider instead of VirtualBox to build your swarm, it is fairly easy to switch the Docker Machine driver and change a few configuration items in the vms_create.sh script (see Provisioning, below).
Required Software
If you want to follow along with this post, you will need the latest versions of git, Docker, Docker Machine, Docker Compose, and VirtualBox installed.
Source Code
All source code for this post is located in two GitHub repositories. The first repository contains scripts to provision the VMs, create an overlay network and persistent host-mounted volumes, build the Docker swarm, and deploy Consul, Registrator, Swarm Visualizer, Fluentd, and the Elastic Stack. The second repository contains scripts to deploy two instances of the Widget Spring Boot RESTful service and a single instance of MongoDB. You can execute all scripts manually, from the command-line, or from a CI/CD pipeline, using tools such as Jenkins.
Provisioning the Swarm
To start, clone the first repository, and execute the single run_all.sh script, or execute the seven individual scripts necessary to provision the VMs, create the overlay network and host volumes, build the swarm, and deploy Consul, Registrator, Swarm Visualizer, Fluentd, and the Elastic Stack. Follow the steps below to complete this part.
When the scripts have completed, the resulting swarm should be configured similarly to the diagram below. Consul, Registrator, Swarm Visualizer, Fluentd, and the Elastic Stack containers should be distributed across the three swarm manager nodes and the three swarm worker nodes (VirtualBox VMs).
Deploying the Application
Next, clone the second repository, and execute the single run_all.sh script, or execute the four scripts necessary to deploy the Widget Spring Boot RESTful service and a single instance of MongoDB. Follow the steps below to complete this part.
When the scripts have completed, the Widget service and MongoDB containers should be distributed across two of the three swarm worker nodes (VirtualBox VMs).
To confirm the final state of the swarm and the running container stacks, use the following Docker commands.
Open the Swarm Visualizer web UI, using any of the swarm manager node IPs, on port 5001, to confirm the swarm health, as well as the running container’s locations.
Lastly, open the Consul Web UI, using any of the swarm manager node IPs, on port 5601, to confirm the running container’s health, as well as their placement on the swarm nodes.
Streaming Logs
Elastic Stack
If you read the previous post, Distributed Service Configuration with Consul, Spring Cloud, and Docker, you will notice we deployed a few additional components this time. First, the Elastic Stack (aka ELK), is deployed to the worker3 swarm worker node, within a single container. I have increased the CPU count and RAM assigned to this VM, to minimally run the Elastic Stack. If you review the docker-compose.yml file, you will note I am using Sébastien Pujadas’ sebp/elk:latest Docker base image from Docker Hub to provision the Elastic Stack. At the time of the post, this was based on the 5.3.0 version of ELK.
Docker Logging Driver
The Widget stack’s docker-compose.yml file has been modified since the last post. The compose file now incorporates a Fluentd logging configuration section for each service. The logging configuration includes the address of the Fluentd instance, on the same swarm worker node. The logging configuration also includes a tag for each log message.
Fluentd
In addition to the Elastic Stack, we have deployed Fluentd to the worker1 and worker2 swarm nodes. This is also where the Widget and MongoDB containers are deployed. Again, looking at the docker-compose.yml file, you will note we are using a custom Fluentd Docker image, garystafford/custom-fluentd:latest, which I created. The custom image is available on Docker Hub.
The custom Fluentd Docker image is based on Fluentd’s official onbuild Docker image, fluent/fluentd:onbuild. Fluentd provides instructions for building your own custom images, from their onbuild base images.
There were two reasons I chose to create a custom Fluentd Docker image. First, I added the Uken Games’ Fluentd Elasticsearch Plugin, to the Docker Image. This highly configurable Fluentd Output Plugin allows us to push Docker logs, processed by Fluentd to the Elasticsearch. Adding additional plugins is a common reason for creating a custom Fluentd Docker image.
The second reason to create a custom Fluentd Docker image was configuration. Instead of bind-mounting host directories or volumes to the multiple Fluentd containers, to provide Fluentd’s configuration, I baked the configuration file into the immutable Docker image. The bare-bones, basicFluentd configuration file defines three processes, which are Input, Filter, and Output. These processes are accomplished using Fluentd plugins. Fluentd has 6 types of plugins: Input, Parser, Filter, Output, Formatter and Buffer. Fluentd is primarily written in Ruby, and its plugins are Ruby gems.
Fluentd listens for input on tcp port 24224, using the forward Input Plugin. Docker logs are streamed locally on each swarm node, from the Widget and MongoDB containers to the local Fluentd container, over tcp port 24224, using Docker’s fluentd logging driver, introduced earlier. Fluentd
Fluentd then filters all input using the stdout Filter Plugin. This plugin prints events to stdout, or logs if launched with daemon mode. This is the most basic method of filtering.
Lastly, Fluentd outputs the filtered input to two destinations, a local log file and Elasticsearch. First, the Docker logs are sent to a local Fluentd log file. This is only for demonstration purposes and debugging. Outputting log files is not recommended for production, nor does it meet the 12-factor application recommendations for logging. Second, Fluentd outputs the Docker logs to Elasticsearch, over tcp port 9200, using the Fluentd Elasticsearch Plugin, introduced above.
Additional Metadata
In addition to the log message itself, in JSON format, the fluentd log driver sends the following metadata in the structured log message: container_id, container_name, and source. This is helpful in identifying and categorizing log messages from multiple sources. Below is a sample of log messages from the raw Fluentd log file, with the metadata tags highlighted in yellow. At the bottom of the output is a log message parsed with jq, for better readability.
Using Elastic Stack
Now that our two Docker stacks are up and running on our swarm, we should be streaming logs to Elasticsearch. To confirm this, open the Kibana web console, which should be available at the IP address of the worker3 swarm worker node, on port 5601.
For the sake of this demonstration, I increased the verbosity of the Spring Boot Widget service’s log level, from INFO to DEBUG, in Consul. At this level of logging, the two Widget services and the single MongoDB instance were generating an average of 250-400 log messages every 30 seconds, according to Kibana.
If that seems like a lot, keep in mind, these are Docker logs, which are single-line log entries. We have not aggregated multi-line messages, such as Java exceptions and stack traces messages, into single entries. That is for another post. Also, the volume of debug-level log messages generated by the communications between the individual services and Consul is fairly verbose.
Inspecting log entries in Kibana, we find the metadata tags contained in the raw Fluentd log output are now searchable fields: container_id, container_name, and source, as well as log. Also, note the _type field, with a value of ‘fluentd’. We injected this field in the output section of our Fluentd configuration, using the Fluentd Elasticsearch Plugin. The _type fiel allows us to differentiate these log entries from other potential data sources.
References
All opinions in this post are my own and not necessarily the views of my current employer or their clients.
Diving Deeper into ‘Getting Started with Spring Cloud’
Posted by Gary A. Stafford in DevOps, Enterprise Software Development, Java Development, Software Development on February 15, 2016
Explore the integration of Spring Cloud and Spring Cloud Netflix tooling, through a deep dive into Pivotal’s ‘Getting Started with Spring Cloud’ presentation.
Introduction
Keeping current with software development and DevOps trends can often make us feel we are, as the overused analogy describes, drinking from a firehose, often several hoses at once. Recently joining a large client engagement, I found it necessary to supplement my knowledge of cloud-native solutions, built with the support of Spring Cloud and Spring Cloud Netflix technologies. One of my favorite sources of information on these subjects is presentations by people like Josh Long, Dr. Dave Syer, and Cornelia Davis of Pivotal Labs, and Jon Schneider and Taylor Wicksell of Netflix.
One presentation, in particular, Getting Started with Spring Cloud, by Long and Syer, provides an excellent end-to-end technical overview of the latest Spring and Netflix technologies. Josh Long’s fast-paced, eighty-minute presentation, available on YouTube, was given at SpringOne2GX 2015 with co-presenter, Dr. Dave Syer, founder of Spring Cloud, Spring Boot, and Spring Batch.
As the presenters of Getting Started with Spring Cloud admit, the purpose of the presentation was to get people excited about Spring Cloud and Netflix technologies, not to provide a deep dive into each technology. However, I believe the presentation’s Reservation Service example provides an excellent learning opportunity. In the following post, we will examine the technologies, components, code, and configuration presented in Getting Started with Spring Cloud. The goal of the post is to provide a greater understanding of the Spring Cloud and Spring Cloud Netflix technologies.
System Overview
Technologies
The presentation’s example introduces a dizzying array of technologies, which include:
Spring Boot
Stand-alone, production-grade Spring-based applications
Spring Data REST / Spring HATEOAS
Spring-based applications following HATEOAS principles
Spring Cloud Config
Centralized external configuration management, backed by Git
Netflix Eureka
REST-based service discovery and registration for failover and load-balancing
Netflix Ribbon
IPC library with built-in client-side software load-balancers
Netflix Zuul
Dynamic routing, monitoring, resiliency, security, and more
Netflix Hystrix
Latency and fault tolerance for distributed system
Netflix Hystrix Dashboard
Web-based UI for monitoring Hystrix
Spring Cloud Stream
Messaging microservices, backed by Redis
Spring Data Redis
Configuration and access to Redis from a Spring app, using Jedis
Spring Cloud Sleuth
Distributed tracing solution for Spring Cloud, sends traces via Thrift to the Zipkin collector service
Twitter Zipkin
Distributed tracing system, backed by Apache Cassandra
H2
In-memory Java SQL database, embedded and server modes
Docker
Package applications with dependencies into standardized Linux containers
System Components
Several components and component sub-systems comprise the presentation’s overall Reservation Service example. Each component implements a combination of the technologies mentioned above. Below is a high-level architectural diagram of the presentation’s example. It includes a few additional features, added as part of this post.
Individual system components include:
Spring Cloud Config Server
Stand-alone Spring Boot application provides centralized external configuration to multiple Reservation system components
Spring Cloud Config Git Repo
Git repository containing multiple Reservation system components configuration files, served by Spring Cloud Config Server
H2 Java SQL Database Server (New)
This post substitutes the original example’s use of H2’s embedded version with a TCP Server instance, shared by Reservation Service instances
Reservation Service
Multi load-balanced instances of stand-alone Spring Boot application, backed by H2 database
Reservation Client
Stand-alone Spring Boot application (aka edge service or client-side proxy), forwards client-side load-balanced requests to the Reservation Service, using Eureka, Zuul, and Ribbon
Reservation Data Seeder (New)
Stand-alone Spring Boot application, seeds H2 with initial data, instead of the Reservation Service
Eureka Service
Stand-alone Spring Boot application provides service discovery and registration for failover and load-balancing
Hystrix Dashboard
Stand-alone Spring Boot application provides web-based Hystrix UI for monitoring system performance and Hystrix circuit-breakers
Zipkin
Zipkin Collector, Query, and Web, and Cassandra database, receives, correlates, and displays traces from Spring Cloud Sleuth
Redis
In-memory data structure store, acting as message broker/transport for Spring Cloud Stream
Github
All the code for this post is available on Github, split between two repositories. The first repository, spring-cloud-demo, contains the source code for all of the components listed above, except the Spring Cloud Config Git Repo. To function correctly, the configuration files, consumed by the Spring Cloud Config Server, needs to be placed into a separate repository, spring-cloud-demo-config-repo.
The first repository contains a git submodule , docker-zipkin. If you are not familiar with submodules, you may want to take a moment to read the git documentation. The submodule contains a dockerized version of Twitter’s OpenZipkin, docker-zipkin. To clone the two repositories, use the following commands. The --recursive option is required to include the docker-zipkin submodule in the project.
Configuration
To try out the post’s Reservation system example, you need to configure at least one property. The Spring Cloud Config Server needs to know the location of the Spring Cloud Config Repository, which is the second GitHub repository you cloned, spring-cloud-demo-config-repo. From the root of the spring-cloud-demo repo, edit the Spring Cloud Config Server application.properties file, located in config-server/src/main/resources/application.properties. Change the following property’s value to your local path to the spring-cloud-demo-config-repo repository:
Startup
There are a few ways you could run the multiple components that make up the post’s example. I suggest running one component per terminal window, in the foreground. In this way, you can monitor the output from the bootstrap and startup processes of the system’s components. Furthermore, you can continue to monitor the system’s components once they are up and running, and receiving traffic. Yes, that is twelve terminal windows…
There is a required startup order for the components. For example, Spring Cloud Config Server needs to start before the other components that rely on it for configuration. Netflix’s Eureka needs to start before the Reservation Client and ReservationServices, so they can register with Eureka on startup. Similarly, Zipkin needs to be started in its Docker container before the Reservation Client and Services, so Spring Cloud Sleuth can start sending traces. Redis needs to be started in its Docker container before Spring Cloud Stream tries to create the message queue. All instances of the Reservation Service needs to start before the Reservation Client. Once every component is started, the Reservation Data Seeder needs to be run once to create initial data in H2. For best results, follow the instructions below. Let each component start completely, before starting the next component.
Docker
Both Zipkin and Redis run in Docker containers. Redis runs in a single container. Zipkin’s four separate components run in four separate containers. Be advised, Zipkin seems to have trouble successfully starting all four of its components on a consistent basis. I believe it’s a race condition caused by Docker Compose simultaneously starting the four Docker containers, ignoring a proper startup order. More than half of the time, I have to stop Zipkin and rerun the docker command to get Zipkin to start without any errors.
If you’ve followed the instructions above, you should see the following Docker images and Docker containers installed and running in your local environment.
Components
Spring Cloud Config Server
At the center of the Reservation system is Spring Cloud Config. Configuration, typically found in the application.properties file, for the Reservation Services, Reservation Client, Reservation Data Seeder, Eureka Service, and Hystix Dashboard, has been externalized with Spring Cloud Config.
Each component has a bootstrap.properties file, which modifies its startup behavior during the bootstrap phase of an application context. Each bootstrap.properties file contains the component’s name and the address of the Spring Cloud Config Server. Components retrieve their configuration from the Spring Cloud Config Server at runtime. Below, is an example of the Reservation Client’s bootstrap.properties file.
Spring Cloud Config Git Repo
In the presentation, as in this post, the Spring Cloud Config Server is backed by a locally cloned Git repository, the Spring Cloud Config Git Repo. The Spring Cloud Config Server’s application.properties file contains the address of the Git repository. Each properties file within the Git repository corresponds to a system component. Below, is an example of the reservation-client.properties file, from the Spring Cloud Config Git Repo.
As shown in the original presentation, the configuration files can be viewed using HTTP endpoints of the Spring Cloud Config Server. To view the Reservation Service’s configuration stored in the Spring Cloud Config Git Repo, issue an HTTP GET request to http://localhost:8888/reservation-service/master. The master URI refers to the Git repo branch in which the configuration resides. This will return the configuration, in the response body, as JSON:
In a real Production environment, the Spring Cloud Config Server would be backed by a highly-available Git Server or GitHub repository.
Reservation Service
The Reservation Service is the core component in the presentation’s example. The Reservation Service is a stand-alone Spring Boot application. By implementing Spring Data REST and Spring HATEOAS, Spring automatically creates REST representations from the Reservation JPA Entity class of the Reservation Service. There is no need to write a Spring Rest Controller and explicitly code each endpoint.
Spring HATEOAS allows us to interact with the Reservation Entity, using HTTP methods, such as GET and POST. These endpoints, along with all addressable endpoints, are displayed in the terminal output when a Spring Boot application starts. For example, we can use an HTTP GET request to call the reservations/{id} endpoint, such as:
The Reservation Service also makes use of the Spring RepositoryRestResource annotation. By annotating the RepositoryReservation Interface, which extends JpaRepository, we can customize export mapping and relative paths of the Reservation JPA Entity class. As shown below, the RepositoryReservation Interface contains the findByReservationName method signature, annotated with /by-name endpoint, which accepts the rn input parameter.
Calling the findByReservationName method, we can search for a particular reservation by using an HTTP GET request to call the reservations/search/by-name?rn={reservationName} endpoint.
Reservation Client
Querying the Reservation Service directly is possible, however, is not the recommended. Instead, the presentation suggests using the Reservation Client as a proxy to the Reservation Service. The presentation offers three examples of using the Reservation Client as a proxy.
The first demonstration of the Reservation Client uses the /message endpoint on the Reservation Client to return a string from the Reservation Service. The message example has been modified to include two new endpoints on the Reservation Client. The first endpoint, /reservations/client-message, returns a message directly from the Reservation Client. The second endpoint, /reservations/service-message, returns a message indirectly from the Reservation Service. To retrieve the message from the Reservation Service, the Reservation Client sends a request to the endpoint Reservation Service’s /message endpoint.
To retrieve both messages, send separate HTTP GET requests to each endpoint:
The second demonstration of the Reservation Client uses a Data Transfer Object (DTO). Calling the Reservation Client’s reservations/names endpoint, invokes the getReservationNames method. This method, in turn, calls the Reservation Service’s /reservations endpoint. The response object returned from the Reservation Service, a JSON array of reservation records, is deserialized and mapped to the Reservation Client’s Reservation DTO. Finally, the method returns a collection of strings, representing just the names from the reservations.
To retrieve the collection of reservation names, an HTTP GET request is sent to the /reservations/names endpoint:
Spring Cloud Stream
One of the more interesting technologies in the presentation is Spring’s Spring Cloud Stream. The Spring website describes Spring Cloud Stream as a project that allows users to develop and run messaging microservices using Spring Integration. In other words, it provides native Spring messaging capabilities, backed by a choice of message buses, including Redis, RabbitMQ, and Apache Kafka, to Spring Boot applications.
A detailed explanation of Spring Cloud Stream would take an entire post. The best technical demonstration I have found is the presentation, Message Driven Microservices in the Cloud, by speakers Dr. David Syer and Dr. Mark Pollack, given in January 2016, also at SpringOne2GX 2015.
In the presentation, a new reservation is submitted via an HTTP POST to the acceptNewReservations method of the Reservation Client. The method, in turn, builds (aka produces) a message, containing the new reservation, and publishes that message to the queue.reservation queue.
The queue.reservation queue is located in Redis, which is running inside a Docker container. To view the messages being published to the queue in real-time, use the redis-cli, with the monitor command, from within the Redis Docker container. Below is an example of tests messages pushed (LPUSH) to the reservations queue from the Reservation Client.
The published messages are consumed by subscribers to the reservation queue. In this example, the consumer is the Reservation Service. The Reservation Service’s acceptNewReservation method processes the message and saves the new reservation to the H2 database. In Spring Cloud Stream terms, the Reservation Client is the Sink.
Netflix Eureka
Netflix’s Eureka, in combination with Netflix’s Zuul and Ribbon, provide the ability to scale the Reservation Service horizontally, and to load balance those instances. By using the @EnableEurekaClient annotation on the Reservation Client and Reservation Services, each instance will automatically register with Eureka on startup, as shown in the Eureka Web UI, below.
The names of the registered instances are in three parts: the address of the host on which the instance is running, followed by the value of the spring.application.name property of the instance’s bootstrap.properties file, and finally, the port number the instance is running on. Eureka displays each instance’s status, along with additional AWS information, if you are running on AWS, as Netflix does.
According to Spring in their informative post, Spring Cloud, service discovery is one of the key tenets of a microservice based architecture. Trying to hand-configure each client, or to rely on convention over configuration, can be difficult to do and is brittle. Eureka is the Netflix Service Discovery Server and Client. A client (Spring Boot application), registers with Eureka, providing metadata about itself. Eureka then receives heartbeat messages from each instance. If the heartbeat fails over a configurable timetable, the instance is normally removed from the registry.
The Reservation Client application is also annotated with @EnableZuulProxy. Adding this annotation pulls in Spring Cloud’s embedded Zuul proxy. Again, according to Spring, the proxy is used by front-end applications to proxy calls to one or more back-end services, avoiding the need to manage CORS and authentication concerns independently for all the backends. In the presentation and this post, the front end is the Reservation Client and the back end is the Reservation Service.
In the code snippet below from the ReservationApiGatewayRestController, note the URL of the endpoint requested in the getReservationNames method. Instead of directly calling http://localhost:8000/reservations, the method calls http://reservation-service/reservations. The reservation-service segment of the URL is the registered name of the service in Eureka and contained in the Reservation Service’s bootstrap.properties file.
In the following abridged output from the Reservation Client, you can clearly see the interaction of Zuul, Ribbon, Eureka, and Spring Cloud Config. Note the Client application has successfully registering itself with Eureka, along with the Reservation Client’s status. Also, note Zuul mapping the Reservation Service’s URL path.
Load Balancing
One shortcoming of the original presentation was true load balancing. With only a single instance of the Reservation Service in the original presentation, there is nothing to load balance; it’s more of a reverse proxy example. To demonstrate load balancing, we need to spin up additional instances of the Reservation Service. Following the post’s component start-up instructions, we should have three instances of the Reservation Service running, on ports 8000, 8001, and 8002, each in separate terminal windows.
To confirm the three instances of the Reservation Service were successfully registered with Eureka, review the output from the Eureka Server terminal window. The output should show three instances of the Reservation Service registering on startup, in addition to the Reservation Client.
Viewing Eureka’s web console, we should observe three members in the pool of Reservation Services.
Lastly, looking at the terminal output of the Reservation Client, we should see three instances of the Reservation Service being returned by Ribbon (aka the DynamicServerListLoadBalancer).
Requesting
Requesting http://localhost:8050/reservations/names, Ribbon forwards the request to one of the three Reservation Service instances registered with Eureka. By default, Ribbon uses a round-robin load-balancing strategy to select an instance from the pool of available Reservation Services.
H2 Server
The original presentation’s Reservation Service used an embedded instance of H2. To scale out the Reservation Service, we need a common database for multiple instances to share. Otherwise, queries would return different results, specific to the particular instance of Reservation Service chosen by the load-balancer. To solve this, the original presentation’s embedded version of H2 has been replaced with the TCP Server client/server version of H2.
Thanks to more Spring magic, the only change we need to make to the original presentation’s code is a few additional properties added to the Reservation Service’s reservation-service.properties file. This changes H2 from the embedded version to the TCP Server version.
Reservation Data Seeder
In the original presentation, the Reservation Service created several sample reservation records in its embedded H2 database on startup. Since we now have multiple instances of the Reservation Service running, the sample data creation task has been moved from the Reservation Service to the new Reservation Data Seeder. The Reservation Service only now validates the H2 database schema on startup. The Reservation Data Seeder now updates the schema based on its entities. This also means the seed data will be persisted across restarts of the Reservation Service, unlike in the original configuration.
Running the Reservation Data Seeder once will create several reservation records into the H2 database. To confirm the H2 Server is running and the initial reservation records were created by the Reservation Data Seeder, point your web browser to the H2 login page at http://192.168.99.1:6889. and log in using the credentials in the reservation-service.properties file.
The H2 Console should contain the RESERVATION table, which holds the reservation sample records.
Spring Cloud Sleuth and Twitter’s Zipkin
According to the project description, “Spring Cloud Sleuth implements a distributed tracing solution for Spring Cloud. All your interactions with external systems should be instrumented automatically. You can capture data simply in logs, or by sending it to a remote collector service.” In our case, that remote collector service is Zipkin.
Zipkin describes itself as, “a distributed tracing system. It helps gather timing data needed to troubleshoot latency problems in microservice architectures. It manages both the collection and lookup of this data through a Collector and a Query service.” Zipkin provides critical insights into how microservices perform in a distributed system.
In the presentation, as in this post, the Reservation Client’s main ReservationClientApplication class contains the alwaysSampler bean, which returns a new instance of org.springframework.cloud.sleuth.sampler.AlwaysSampler. As long as Spring Cloud Sleuth is on the classpath and you have added alwaysSampler bean, the Reservation Client will automatically generate trace data.
Sending a request to the Reservation Client’s service/message endpoint (http://localhost:8050/reservations/service-message,), will generate a trace, composed of spans. in this case, the spans are individual segments of the HTTP request/response lifecycle. Traces are sent by Sleuth to Zipkin, to be collected. According to Spring, if spring-cloud-sleuth-zipkin is available, then the application will generate and collect Zipkin-compatible traces using Brave). By default, it sends them via Apache Thrift to a Zipkin collector service on port 9410.
Zipkin’s web-browser interface, running on port 8080, allows us to view traces and drill down into individual spans.
Zipkin contains fine-grain details about each span within a trace, as shown below.
Correlation IDs
Note the x-trace-id and x-span-id in the request header, shown below. Sleuth injects the trace and span IDs to the SLF4J MDC (Simple Logging Facade for Java – Mapped Diagnostic Context). According to Spring, IDs provides the ability to extract all the logs from a given trace or span in a log aggregator. The use of correlation IDs and log aggregation are essential for monitoring and supporting a microservice architecture.
Hystix and Hystrix Dashboard
The last major technology highlighted in the presentation is Netflix’s Hystrix. According to Netflix, “Hystrix is a latency and fault tolerance library designed to isolate points of access to remote systems, services, and 3rd party libraries, stop cascading failure and enable resilience in complex distributed systems where failure is inevitable.” Hystrix is essential, it protects applications from cascading dependency failures, an issue common to complex distributed architectures, with multiple dependency chains. According to Netflix, Hystrix uses multiple isolation techniques, such as bulkhead, swimlane, and circuit breaker patterns, to limit the impact of any one dependency on the entire system.
The presentation demonstrates one of the simpler capabilities of Hystrix, fallback. The getReservationNames method is decorated with the @HystrixCommand annotation. This annotation contains the fallbackMethod. According to Netflix, a graceful degradation of a method is provided by adding a fallback method. Hystrix will call to obtain a default value or values, in case the main command fails. In the presentation’s example, the Reservation Service, a direct dependency of the Reservation Client, has failed. The Reservation Service failure causes the failure of the Reservation Client.
In the presentation’s example, the Reservation Service, a direct dependency of the Reservation Client, has failed. The Reservation Service failure causes the failure of the Reservation Client’s getReservationNames method to return a collection of reservation names. Hystrix redirects the application to the getReservationNameFallback method. Instead of returning a collection of reservation names, the getReservationNameFallback returns an empty collection, as opposed to an error message to the client.
A more relevant example involves Netflix movie recommendation service. In the event a failure of the recommendation service’s method to return a collection of personalized list of movie recommendations to a customer, Hystrix fallbacks to a method that returns a generic list of the most popular movies to the customer. Netflix has determined that, in the event of a failure of their recommendation service, falling back to a generic list of movies is better than returning no movies at all.
The Hystrix Dashboard is a tool, available with Hystrix, to visualize the current state of Hystrix instrumented methods. Although visually simplistic, the dashboard effectively presents the health of calls to external systems, which are wrapped in a HystrixCommand or HystrixObservableCommand.
The Hystrix dashboard is a visual representation of the Hystrix Stream. This stream is a live feed of data sent by the Hystrix instrumented application, in this case, the Reservation Client. For a single Hystrix application, such as the Reservation Client, the feed requested from the application’s hystrix.stream endpoint is http://localhost:8050/hystrix.stream. The dashboard consumes the stream resource’s response and visualizes it in the browser using JavaScript, jQuery, and d3.
In the post, as in the presentation, hitting the Reservation Client with a volume of requests, we observe normal activity in Hystrix Dashboard. All three instances of the Reservation Service are running and returning the collection of reservations from H2, to the Reservation Client.
If all three instances of the Reservation Service fail or the maximum latency is exceeded, the Reservation Client falls back to returning an empty collection in the response body. In the example below, 15 requests, representing 100% of the current traffic, to the getReservationNames method failed and subsequently fell back to return an empty collection. Hystrix succeeded in helping the application gracefully fall back to an alternate response.
Conclusion
It’s easy to see how Spring Cloud and Netflix’s technologies are easily combined to create a performant, horizontally scalable, reliable system. With the addition of a few missing components, such metrics monitoring and log aggregation, this example could easily be scaled up to support a production-grade microservices-based, enterprise software platform.
Gary Stafford
Accenture | Technology Architecture Delivery | Senior Manager | DevOps | Software Development | Docker | Microservices | JavaScript | Java | IaC | AWS
