hoang-ho / Microservices-Web-App

Microservices Web App built with Python Flask

Multi-tier Microservices WebApp

Programming Lab2 for the course CS 677: Distributed OS. In this lab we have to implement a microservices architecture using REST APIs. We have to emulate a bookstore running on a single server with a single buyer. The bookstore components (front-end, catalogue and order) are deployed as three processes and a fourth process representing the client performing interface calls to the front-end process.

Assumption: As per the documentation, the search operation returns only id and topic fields and the lookup operation returns only cost and stock fields.

To deploy to EC2

Creating and Running Instances

Create 3 EC2 instances using ami-085ca2d6000e2a2e9 (this is a customized image with docker, docker-composed and git repo cloned and installed).

NOTE: If you want to use your own ami, you will need to set up docker, git and security group beforehand

$ aws ec2 run-instances --image-id ami-085ca2d6000e2a2e9 --instance-type t2.micro --key-name 677kp
$ aws ec2 describe-instances --instance-id $INSTANCE_ID

From the last command, save down the Public IPv4 DNS and the Private IPv4 addresses for each instance. Next, ssh into the instance

Updating the Security Group Settings

Confirm the rules for incoming traffic as follows. In the Inbound rules section, create the following rules (choose Add rule for each new rule):

• Choose HTTP from the Type list, port 80 for port range, and make sure that Source is set to Anywhere (0.0.0.0/0).
• Choose HTTPS from the Type list, port 443 for port range, and make sure that Source is set to Anywhere (0.0.0.0/0).
• Choose Custom TCP from the Type list, 5000-6000 for Port Range, and make sure that Source is set to Anywhere (0.0.0.0/0).

Setting up the Instances

ssh into each of the instance as follows:

$ ssh -i <path-to-677kp.pem-file> ec2-user@$PUBLIC_IPv4_DNS

Change the working directory to the repository's root folder

$ cd Microservices-Web-App

Setting up the config_env File

For the instance you choose to server the catalog-service, put its Private IPv4 address in the ./config_env file for the field name CATALOG_HOST, similarly update the ORDER_HOST for the order-service.

You repeat this step for EC2 instances that you choose to server front-end-service and the order-service as follows:

NOTE: If you want to change the port, please remember to update the security group, the config_env and the docker-compose file!

# Modify the config_env
$ vim config_env

Deploying and Running Containers in EC2 Instances

Now we're done setting up and will start deploying as follows:

In the instance you choose to be front-end-service run

$ docker-compose up --build front-end-service

In the instance you choose to be catalog-service run

$ docker-compose up --build catalog-service

In the instance you choose to be order-service run

$ docker-compose up --build order-service

Testing from the Client

From your local machine, to test the API run the following commands

$ curl --request GET $FRONT_END_PUBLIC_IPv4_DNS:5004/search/<topic_name>

$ curl --request POST $FRONT_END_PUBLIC_IPv4_DNS:5004/buy/<book_id>

$ curl --request GET $FRONT_END_PUBLIC_IPv4_DNS:5004/lookup/<book_id>

curl --header "Content-Type: application/json" --request PUT  --data '{"id": 1, "stock":2000, "cost":2000}' http://$CATALOG_PUBLIC_IPv4_DNS:5002/catalog/update

Sample Output:

NOTE: More sample outputs can be found under ./doc/outputs folder

curl --request GET $FRONT_END_PUBLIC_IPv4_DNS:5004/search/distributed-systems
{
    "items": {
        "How to get a good grade in 677 in 20 minutes a day.": 1,
        "RPCs for Dummies.": 2
    }
}

Test Scripts for Testing

1. To test services deployed on AWS EC2 instances

   Follow the To Deploy to EC2 Instance steps and then run the below command from the client that has the git repo cloned. The output of the test is in a newly created txt file - out.txt.

   $ cd Microservices-Web-App
   $ bash test.sh $FRONT_END_SERVER_PUBLIC_IPv4_DNS $CATALOG_SERVER_PUBLIC_IPv4_DNS
   

2. To test services deployed locally

   Follow the To Deploy to EC2 Instance steps and then run the below command from the client that has the git repo cloned. The output of the test is in a newly created txt file - out.txt.

   NOTE: Please don't update the ./config_env file on the client machine to run the tests locally.

   $ cd Microservices-Web-App
   $ bash test_local.sh
   

Simulating Concurrency

To simulate a concurrency situation with buy and update, run the python file ./SimulateConcurrency.py.

NOTE: Please make sure to have the requests library installed on the client.

Sample Output:

(base) Hoangs-MacBook-Pro:Microservices-Web-App hoangho$ python3 SimulateConcurrency.py --front-end-dns $FRONT_END_SERVER_PUBLIC_IPv4_DNS  --catalog-dns $CATALOG_SERVER_PUBLIC_IPv4_DNS 

INFO:root:Look up the book stock and cost before update and buy: {
    "cost": 10.0,
    "stock": 1000
}
INFO:root:Main    : create and start thread 0.
INFO:root:Calling request http://ec2-100-25-36-171.compute-1.amazonaws.com/buy/2 at timestamp 1617657154.635759
INFO:root:Main    : create and start thread 1.
INFO:root:Calling request http://ec2-100-25-36-171.compute-1.amazonaws.com/buy/2 at timestamp 1617657154.6362062
INFO:root:Main    : create and start thread 2.
INFO:root:Calling request http://ec2-54-210-80-160.compute-1.amazonaws.com/catalog/update at timestamp 1617657154.636657
INFO:root:Calling request http://ec2-54-210-80-160.compute-1.amazonaws.com/catalog/update at timestamp 1617657154.636717
INFO:root:Main    : before joining thread 0.
INFO:root:Response: {
  "book": "RPCs for Dummies.", 
  "message": "Done update"
}
 at time stamp 1617657154.725301
INFO:root:Response: {
    "message": "successfully purchased the book RPCs for Dummies."
}
 at time stamp 1617657154.74664
INFO:root:Main    : thread 0 done
INFO:root:Main    : before joining thread 1.
INFO:root:Response: {
    "message": "successfully purchased the book RPCs for Dummies."
}
 at time stamp 1617657154.758931
INFO:root:Main    : thread 1 done
INFO:root:Main    : before joining thread 2.
INFO:root:Main    : thread 2 done
INFO:root:Look up the book stock and cost after update and buy: {
    "cost": 2000.0,
    "stock": 1998
}

As we can see from the terminal log, we have 2 update requests and 1 buy requests. The two update requests are done first, and thus, the stock for book 2 is set to 2000.0 and its cost is set to 2000.0. When the buy request comes in, the stock is set to 1999 and the cost remains the same!

Evalation

1. Time to complete 1000 sequential requests by one user

INFO:root:Total time: 90191.7736530304
INFO:root:Average time: 90.1917736530304

2. Time to complete 1000 sequential requests by 5 user at a time

INFO:root:Total time: 136533.04195404053
INFO:root:Average time: 136.53304195404053

Logging on Catalog and Order Services

1. Catalog Service

   Logging happens in logfile.json inside the Docker container catalog-service. To check the logs run the following command while the container is running

   $ docker exec -it catalog-service bash 
   

   Inside the container, check the logfile.json:

   $ cat logfile.json
   

   A buy request will be logged in the "buy" key of the json object.

   A query request will be logged in the "query" key of the json object.

2. Order Service

   To view the logs for the order-service call the following request while the container is running

   curl --header "Content-Type: application/json" --request GET http://$ORDER_PUBLIC_IPv4_DNS:5007/log