Today, i learnt about Rate Limiting Using Redis which is very strange for me. In many blogs, they mentioned implementing redis as a rate limiter, and the code is written in server side. Shouldn’t be offloaded to the gateway ? Many such questions arised. In the mean time, i re-learned the rate limit concept. In this blog i jot down notes on rate limiting with examples using HAProxy (can be used in production).

What is Rate Limiting?

Rate limiting refers to the process of limiting the number of requests a user, application, or IP address can make to a system in a given period. This mechanism is essential to protect systems from:

• Overloading caused by high traffic or malicious activity.
• Denial of Service (DoS) attacks.
• Resource starvation due to unbalanced usage patterns.
• Ensuring fair usage among all clients.

For example, a public API might limit each user to 100 requests per minute to avoid overwhelming the backend systems.

Types of Rate Limiting

Rate limiting can be implemented in various ways depending on the use case and infrastructure. Here are the most common types

1. Fixed Window
2. Sliding Window
3. Token Bucket
4. Leaky Bucket
5. Concurrent Rate Limiting

1. Fixed Window Rate Limiting

In this method, a fixed time window (e.g., 1 minute) is defined, and a request counter is maintained. If the number of requests exceeds the allowed limit within the window, subsequent requests are denied.

How It Works

• A counter is initialized at the start of the time window.
• Each incoming request increments the counter.
• If the counter exceeds the predefined limit, the request is rejected until the window resets.

Advantages

• Simple to implement.
• Effective for scenarios where traffic is predictable.

Disadvantages

• Burst traffic at the boundary of two windows can lead to uneven load, as a user can send the maximum requests at the end of one window and immediately at the start of the next.

Example: Allow 60 requests per minute. At the start of each new minute, the counter resets.

Implementation:

HAProxy offers various methods to control traffic, including rate limiting. We can implement fixed window rate limiting in HAProxy using stick tables, which are in-memory tables used to store information about each connection. These stick tables can be configured to store request counts, track IP addresses, and enforce rate limits.

Step 1: Define the Stick Table

To track the number of requests for a given client, we define a stick table that holds the request count and sets the expiration for the time window.

backend my_backend
    stick-table type ip size 1m expire 60s store http_req_rate(60s)

Explanation:

• type ip: This means that the stick table will track client IP addresses.
• size 1m: This defines the size of the table (1 million entries in this case).
• expire 60s: The table will expire every 60 seconds (i.e., every fixed time window).
• store http_req_rate(60s): This stores the request rate per IP over the last 60 seconds.

Step 2: Apply Rate Limiting Based on the Stick Table

Next, you apply rate limiting based on the values stored in the stick table. You can reject requests that exceed the allowed rate limit by using the http-request directive.

frontend http_in
    bind *:80
    acl too_many_requests sc_http_req_rate(my_backend) gt 100
    http-request deny if too_many_requests
    default_backend my_backend

Explanation:

• acl too_many_requests sc_http_req_rate(my_backend) gt 100: This defines an Access Control List (ACL) that checks if the request rate for a particular IP (from the stick table) exceeds 100 requests in the last 60 seconds.
• http-request deny if too_many_requests: If the ACL condition is met (i.e., the IP exceeds the rate limit), the request is denied.

2. Sliding Window Rate Limiting

This approach improves upon fixed windows by using a rolling window. Requests are tracked using timestamps, and the rate is calculated based on a dynamic window.

How It Works

• Each request is timestamped.
• A sliding window keeps track of all requests within a defined time frame.
• The system calculates the total requests in the window to determine whether a new request is allowed.

Advantages

• Reduces the impact of bursts near window boundaries.
• Provides a smoother throttling experience.

Disadvantages

• Slightly more complex to implement due to the need for maintaining and evaluating timestamps.

Example: Allow 60 requests over the last 60 seconds, calculated dynamically.

Implementation:

In this scenario, you want to limit the number of requests that a user can make within a certain period of time. The period is a sliding window. So, if you set it to allow no more than 20 requests per client during the last 10 seconds, HAProxy will count the last 10 seconds. Consider this HAProxy configuration,

frontend website
    bind :80
    stick-table  type ipv6  size 100k  expire 30s  store http_req_rate(10s)
    http-request track-sc0 src
    http-request deny deny_status 429 if { sc_http_req_rate(0) gt 20 }
    default_backend servers

The stick-table directive in HAProxy creates a key-value store to track counters like HTTP request rates per client. The client’s IP address is used as the key, and its request count is stored and aggregated. The http-request track-sc0 src line adds the client to the stick table, starting the count of their requests.

Records in the stick table expire after a specified inactivity period, as defined by the expire parameter, which helps free up space. Without an expire parameter, the oldest records are evicted once the table is full. For example, 100,000 records can be allowed.

The http-request deny line enforces the rate limit and specifies the action when the limit is exceeded. In this case, the rate is limited to 20 concurrent requests, and any additional requests are denied with a 429 status code until the count falls below the threshold. Other actions can include forwarding to a dedicated backend or silently dropping the connection. The sc_http_req_rate method fetches the client’s current request rate.

You can adjust the time period or threshold, such as allowing up to 1000 requests over 24 hours by changing http_req_rate(10s) to http_req_rate(24h) and updating the deny line accordingly.

3. Token Bucket Algorithm

This algorithm uses a bucket to hold tokens, where each token represents a request. Tokens are replenished at a fixed rate. A request is processed only if a token is available; otherwise, it is rejected or delayed.

How It Works:

• A bucket holds a maximum number of tokens (capacity).
• Tokens are added to the bucket at a steady rate.
• When a request is received, a token is removed from the bucket.
• If the bucket is empty, the request is rejected or delayed until a token becomes available.

Advantages:

• Allows for short bursts of activity while maintaining overall limits.
• Efficient and widely used.

Disadvantages:

• Complex to set up in distributed systems.

Example: Refill 10 tokens per second, with a maximum bucket capacity of 100 tokens.

4. Leaky Bucket Algorithm

Similar to the token bucket but focuses on maintaining a consistent outflow of requests. Excess requests are queued and processed at a steady rate.

How It Works:

• Requests enter a queue (bucket).
• The system processes requests at a fixed rate.
• If the queue is full, additional requests are rejected or delayed.

Advantages:

• Ensures a constant request rate.
• Good for smoothing out traffic bursts.

Disadvantages:

• May introduce latency due to queuing.

Example: Process requests at a steady rate of 5 per second, regardless of the input rate.

5. Concurrent Rate Limiting

Limits the number of concurrent requests a user or system can make.

How It Works:

• The system tracks the number of active or ongoing requests for each user.
• If the active requests exceed the limit, new requests are rejected until one or more ongoing requests are completed.

Advantages:

• Useful for systems with high latency or long-lived connections.
• Prevents resource exhaustion from simultaneous requests.

Disadvantages:

• May require complex state management to track active requests.

Example: Allow a maximum of 10 simultaneous requests per user.

References:

1. https://www.haproxy.com/blog/four-examples-of-haproxy-rate-limiting
2. https://medium.com/@m-elbably/rate-limiting-a-dynamic-distributed-rate-limiting-with-redis-339f9504200f
3. https://irshitmukherjee55.hashnode.dev/rate-limiting-using-redis-golang-token-bucket-algorithm
4. https://www.infoworld.com/article/2257527/how-to-use-redis-for-real-time-metering-applications.html
5. https://docs.aws.amazon.com/waf/latest/developerguide/waf-rule-statement-type-rate-based-request-limiting.html

Load balancing helps distribute client requests across multiple servers to ensure high availability, performance, and reliability. Weighted Round Robin Load Balancing is an extension of the round-robin algorithm, where each server is assigned a weight based on its capacity or performance capabilities. This approach ensures that more powerful servers handle more traffic, resulting in a more efficient distribution of the load.

What is Weighted Round Robin Load Balancing?

Weighted Round Robin Load Balancing assigns a weight to each server. The weight determines how many requests each server should handle relative to the others. Servers with higher weights receive more requests compared to those with lower weights. This method is useful when backend servers have different processing capabilities or resources.

Step-by-Step Implementation with Docker

Step 1: Create Dockerfiles for Each Flask Application

We’ll use the same three Flask applications (app1.py, app2.py, and app3.py) as in previous examples.

• Flask App 1 (app1.py):

from flask import Flask

app = Flask(__name__)

@app.route("/")
def home():
    return "Hello from Flask App 1!"

@app.route("/data")
def data():
    return "Data from Flask App 1!"

if __name__ == "__main__":
    app.run(host="0.0.0.0", port=5001)

• Flask App 2 (app2.py):

from flask import Flask

app = Flask(__name__)

@app.route("/")
def home():
    return "Hello from Flask App 2!"

@app.route("/data")
def data():
    return "Data from Flask App 2!"

if __name__ == "__main__":
    app.run(host="0.0.0.0", port=5002)

• Flask App 3 (app3.py):

from flask import Flask

app = Flask(__name__)

@app.route("/")
def home():
    return "Hello from Flask App 3!"

@app.route("/data")
def data():
    return "Data from Flask App 3!"

if __name__ == "__main__":
    app.run(host="0.0.0.0", port=5003)

Step 2: Create Dockerfiles for Each Flask Application

Create Dockerfiles for each of the Flask applications:

• Dockerfile for Flask App 1 (Dockerfile.app1):

# Use the official Python image from Docker Hub
FROM python:3.9-slim

# Set the working directory inside the container
WORKDIR /app

# Copy the application file into the container
COPY app1.py .

# Install Flask inside the container
RUN pip install Flask

# Expose the port the app runs on
EXPOSE 5001

# Run the application
CMD ["python", "app1.py"]

• Dockerfile for Flask App 2 (Dockerfile.app2):

FROM python:3.9-slim
WORKDIR /app
COPY app2.py .
RUN pip install Flask
EXPOSE 5002
CMD ["python", "app2.py"]

• Dockerfile for Flask App 3 (Dockerfile.app3):

FROM python:3.9-slim
WORKDIR /app
COPY app3.py .
RUN pip install Flask
EXPOSE 5003
CMD ["python", "app3.py"]

Step 3: Create the HAProxy Configuration File

Create an HAProxy configuration file (haproxy.cfg) to implement Weighted Round Robin Load Balancing

global
    log stdout format raw local0
    daemon

defaults
    log     global
    mode    http
    option  httplog
    option  dontlognull
    timeout connect 5000ms
    timeout client  50000ms
    timeout server  50000ms

frontend http_front
    bind *:80
    default_backend servers

backend servers
    balance roundrobin
    server server1 app1:5001 weight 2 check
    server server2 app2:5002 weight 1 check
    server server3 app3:5003 weight 3 check

Explanation:

• The balance roundrobin directive tells HAProxy to use the Round Robin load balancing algorithm.
• The weight option for each server specifies the weight associated with each server:
  • server1 (App 1) has a weight of 2.
  • server2 (App 2) has a weight of 1.
  • server3 (App 3) has a weight of 3.
• Requests will be distributed based on these weights: App 3 will receive the most requests, App 2 the least, and App 1 will be in between.

Step 4: Create a Dockerfile for HAProxy

Create a Dockerfile for HAProxy (Dockerfile.haproxy):

# Use the official HAProxy image from Docker Hub
FROM haproxy:latest

# Copy the custom HAProxy configuration file into the container
COPY haproxy.cfg /usr/local/etc/haproxy/haproxy.cfg

# Expose the port for HAProxy
EXPOSE 80

Step 5: Create a docker-compose.yml File

To manage all the containers together, create a docker-compose.yml file

version: '3'

services:
  app1:
    build:
      context: .
      dockerfile: Dockerfile.app1
    container_name: flask_app1
    ports:
      - "5001:5001"

  app2:
    build:
      context: .
      dockerfile: Dockerfile.app2
    container_name: flask_app2
    ports:
      - "5002:5002"

  app3:
    build:
      context: .
      dockerfile: Dockerfile.app3
    container_name: flask_app3
    ports:
      - "5003:5003"

  haproxy:
    build:
      context: .
      dockerfile: Dockerfile.haproxy
    container_name: haproxy
    ports:
      - "80:80"
    depends_on:
      - app1
      - app2
      - app3

Explanation:

• The docker-compose.yml file defines the services (app1, app2, app3, and haproxy) and their respective configurations.
• HAProxy depends on the three Flask applications to be up and running before it starts.

Step 6: Build and Run the Docker Containers

Run the following command to build and start all the containers

docker-compose up --build

This command builds Docker images for all three Flask apps and HAProxy, then starts them.

Step 7: Test the Load Balancer

Open your browser or use curl to make requests to the HAProxy server

curl http://localhost/
curl http://localhost/data

Observation:

• With Weighted Round Robin Load Balancing, you should see that requests are distributed according to the weights specified in the HAProxy configuration.
• For example, App 3 should receive three times more requests than App 2, and App 1 should receive twice as many as App 2.

Conclusion

By implementing Weighted Round Robin Load Balancing with HAProxy, you can distribute traffic more effectively according to the capacity or performance of each backend server. This approach helps optimize resource utilization and ensures a balanced load across servers.

Load balancing distributes client requests across multiple servers to ensure high availability and reliability. One of the simplest load balancing algorithms is Random Load Balancing, which selects a backend server randomly for each client request.

Although this approach does not consider server load or other metrics, it can be effective for less critical applications or when the goal is to achieve simplicity.

What is Random Load Balancing?

Random Load Balancing assigns incoming requests to a randomly chosen server from the available pool of servers. This method is straightforward and ensures that requests are distributed in a non-deterministic manner, which may work well for environments with equally capable servers and minimal concerns about server load or state.

Step-by-Step Implementation with Docker

Step 1: Create Dockerfiles for Each Flask Application

We’ll use the same three Flask applications (app1.py, app2.py, and app3.py) as in previous examples.

Flask App 1 – (app.py)

from flask import Flask

app = Flask(__name__)

@app.route("/")
def home():
    return "Hello from Flask App 1!"

@app.route("/data")
def data():
    return "Data from Flask App 1!"

if __name__ == "__main__":
    app.run(host="0.0.0.0", port=5001)

Flask App 2 – (app.py)

from flask import Flask

app = Flask(__name__)

@app.route("/")
def home():
    return "Hello from Flask App 2!"

@app.route("/data")
def data():
    return "Data from Flask App 2!"

if __name__ == "__main__":
    app.run(host="0.0.0.0", port=5002)

Flask App 3 – (app.py)

from flask import Flask

app = Flask(__name__)

@app.route("/")
def home():
    return "Hello from Flask App 3!"

@app.route("/data")
def data():
    return "Data from Flask App 3!"

if __name__ == "__main__":
    app.run(host="0.0.0.0", port=5003)

Step 2: Create Dockerfiles for Each Flask Application

Create Dockerfiles for each of the Flask applications:

• Dockerfile for Flask App 1 (Dockerfile.app1):

# Use the official Python image from Docker Hub
FROM python:3.9-slim

# Set the working directory inside the container
WORKDIR /app

# Copy the application file into the container
COPY app1.py .

# Install Flask inside the container
RUN pip install Flask

# Expose the port the app runs on
EXPOSE 5001

# Run the application
CMD ["python", "app1.py"]

• Dockerfile for Flask App 2 (Dockerfile.app2):

FROM python:3.9-slim
WORKDIR /app
COPY app2.py .
RUN pip install Flask
EXPOSE 5002
CMD ["python", "app2.py"]

• Dockerfile for Flask App 3 (Dockerfile.app3):

FROM python:3.9-slim
WORKDIR /app
COPY app3.py .
RUN pip install Flask
EXPOSE 5003
CMD ["python", "app3.py"]

Step 3: Create a Dockerfile for HAProxy

HAProxy Configuration file,

global
    log stdout format raw local0
    daemon

defaults
    log     global
    mode    http
    option  httplog
    option  dontlognull
    timeout connect 5000ms
    timeout client  50000ms
    timeout server  50000ms

frontend http_front
    bind *:80
    default_backend servers

backend servers
    balance random
    random draw 2
    server server1 app1:5001 check
    server server2 app2:5002 check
    server server3 app3:5003 check

Explanation:

• The balance random directive tells HAProxy to use the Random load balancing algorithm.
• The random draw 2 setting makes HAProxy select 2 servers randomly and choose the one with the least number of connections. This adds a bit of load awareness to the random choice.
• The server directives define the backend servers and their ports.

Step 4: Create a Dockerfile for HAProxy

Create a Dockerfile for HAProxy (Dockerfile.haproxy):

# Use the official HAProxy image from Docker Hub
FROM haproxy:latest

# Copy the custom HAProxy configuration file into the container
COPY haproxy.cfg /usr/local/etc/haproxy/haproxy.cfg

# Expose the port for HAProxy
EXPOSE 80

Step 5: Create a docker-compose.yml File

To manage all the containers together, create a docker-compose.yml file:

version: '3'

services:
  app1:
    build:
      context: .
      dockerfile: Dockerfile.app1
    container_name: flask_app1
    ports:
      - "5001:5001"

  app2:
    build:
      context: .
      dockerfile: Dockerfile.app2
    container_name: flask_app2
    ports:
      - "5002:5002"

  app3:
    build:
      context: .
      dockerfile: Dockerfile.app3
    container_name: flask_app3
    ports:
      - "5003:5003"

  haproxy:
    build:
      context: .
      dockerfile: Dockerfile.haproxy
    container_name: haproxy
    ports:
      - "80:80"
    depends_on:
      - app1
      - app2
      - app3

Explanation:

• The docker-compose.yml file defines the services (app1, app2, app3, and haproxy) and their respective configurations.
• HAProxy depends on the three Flask applications to be up and running before it starts.

Step 6: Build and Run the Docker Containers

Run the following command to build and start all the containers:

docker-compose up --build

This command builds Docker images for all three Flask apps and HAProxy, then starts them.

Step 7: Test the Load Balancer

Open your browser or use curl to make requests to the HAProxy server:

curl http://localhost/
curl http://localhost/data

Observation:

• With Random Load Balancing, each request should randomly hit one of the three backend servers.
• Since the selection is random, you may not see a predictable pattern; however, the requests should be evenly distributed across the servers over a large number of requests.

Conclusion

By implementing Random Load Balancing with HAProxy, we’ve demonstrated a simple way to distribute traffic across multiple servers without relying on complex metrics or state information. While this approach may not be ideal for all use cases, it can be useful in scenarios where simplicity is more valuable than fine-tuned load distribution.

Load balancing is crucial for distributing incoming network traffic across multiple servers, ensuring optimal resource utilization and improving application performance. One of the simplest and most popular load balancing algorithms is Round Robin. In this blog, we’ll explore how to implement Least Connection load balancing using Flask as our backend application and HAProxy as our load balancer.

What is Least Connection Load Balancing?

Least Connection Load Balancing is a dynamic algorithm that distributes requests to the server with the fewest active connections at any given time. This method ensures that servers with lighter loads receive more requests, preventing any single server from becoming a bottleneck.

Step-by-Step Implementation with Docker

Step 1: Create Dockerfiles for Each Flask Application

We’ll create three separate Dockerfiles, one for each Flask app.

Flask App 1 (app1.py) – Introduced Slowness by adding sleep

from flask import Flask
import time

app = Flask(__name__)

@app.route("/")
def hello():
    time.sleep(5)
    return "Hello from Flask App 1!"

if __name__ == "__main__":
    app.run(host="0.0.0.0", port=5001)

Flask App 2 (app2.py)

from flask import Flask

app = Flask(__name__)

@app.route("/")
def hello():
    return "Hello from Flask App 2!"

if __name__ == "__main__":
    app.run(host="0.0.0.0", port=5002)

Flask App 3 (app3.py) – Introduced Slowness by adding sleep.

from flask import Flask
import time

app = Flask(__name__)

@app.route("/")
def hello():
    time.sleep(5)
    return "Hello from Flask App 3!"

if __name__ == "__main__":
    app.run(host="0.0.0.0", port=5003)

Each Flask app listens on a different port (5001, 5002, 5003).

Step 2: Dockerfiles for each flask application

Dockerfile for Flask App 1 (Dockerfile.app1)

# Use the official Python image from the Docker Hub
FROM python:3.9-slim

# Set the working directory inside the container
WORKDIR /app

# Copy the current directory contents into the container at /app
COPY app1.py .

# Install Flask inside the container
RUN pip install Flask

# Expose the port the app runs on
EXPOSE 5001

# Run the application
CMD ["python", "app1.py"]

Dockerfile for Flask App 2 (Dockerfile.app2)

FROM python:3.9-slim
WORKDIR /app
COPY app2.py .
RUN pip install Flask
EXPOSE 5002
CMD ["python", "app2.py"]

Dockerfile for Flask App 3 (Dockerfile.app3)

FROM python:3.9-slim
WORKDIR /app
COPY app3.py .
RUN pip install Flask
EXPOSE 5003
CMD ["python", "app3.py"]

Step 3: Create a configuration for HAProxy

global
    log stdout format raw local0
    daemon

defaults
    log     global
    mode    http
    option  httplog
    option  dontlognull
    timeout connect 5000ms
    timeout client  50000ms
    timeout server  50000ms

frontend http_front
    bind *:80
    default_backend servers

backend servers
    balance leastconn
    server server1 app1:5001 check
    server server2 app2:5002 check
    server server3 app3:5003 check

Explanation:

• frontend http_front: Defines the entry point for incoming traffic. It listens on port 80.
• backend servers: Specifies the servers HAProxy will distribute traffic evenly the three Flask apps (app1, app2, app3). The balance leastconn directive sets the Least Connection for load balancing.
• server directives: Lists the backend servers with their IP addresses and ports. The check option allows HAProxy to monitor the health of each server.

Step 4: Create a Dockerfile for HAProxy

Create a Dockerfile for HAProxy (Dockerfile.haproxy)

# Use the official HAProxy image from Docker Hub
FROM haproxy:latest

# Copy the custom HAProxy configuration file into the container
COPY haproxy.cfg /usr/local/etc/haproxy/haproxy.cfg

# Expose the port for HAProxy
EXPOSE 80

Step 5: Create a Dockercompose file

To manage all the containers together, create a docker-compose.yml file

version: '3'

services:
  app1:
    build:
      context: .
      dockerfile: Dockerfile.app1
    container_name: flask_app1
    ports:
      - "5001:5001"

  app2:
    build:
      context: .
      dockerfile: Dockerfile.app2
    container_name: flask_app2
    ports:
      - "5002:5002"

  app3:
    build:
      context: .
      dockerfile: Dockerfile.app3
    container_name: flask_app3
    ports:
      - "5003:5003"

  haproxy:
    build:
      context: .
      dockerfile: Dockerfile.haproxy
    container_name: haproxy
    ports:
      - "80:80"
    depends_on:
      - app1
      - app2
      - app3

Explanation:

• The docker-compose.yml file defines four services: app1, app2, app3, and haproxy.
• Each Flask app is built from its respective Dockerfile and runs on its port.
• HAProxy is configured to wait (depends_on) for all three Flask apps to be up and running.

Step 6: Build and Run the Docker Containers

Run the following commands to build and start all the containers:

# Build and run the containers
docker-compose up --build

This command will build Docker images for all three Flask apps and HAProxy and start them up in the background.

You should see the responses alternating between “Hello from Flask App 1!”, “Hello from Flask App 2!”, and “Hello from Flask App 3!” as HAProxy uses the Round Robin algorithm to distribute requests.

Step 7: Test the Load Balancer

Open your browser or use a tool like curl to make requests to the HAProxy server:

curl http://localhost

You should see responses cycling between “Hello from Flask App 1!”, “Hello from Flask App 2!”, and “Hello from Flask App 3!” according to the Least Connection strategy.

Load balancing is crucial for distributing incoming network traffic across multiple servers, ensuring optimal resource utilization and improving application performance. One of the simplest and most popular load balancing algorithms is Round Robin. In this blog, we’ll explore how to implement Round Robin load balancing using Flask as our backend application and HAProxy as our load balancer.

What is Round Robin Load Balancing?

Round Robin load balancing works by distributing incoming requests sequentially across a group of servers.

For example, the first request goes to Server A, the second to Server B, the third to Server C, and so on. Once all servers have received a request, the cycle repeats. This algorithm is simple and works well when all servers have similar capabilities.

Step-by-Step Implementation with Docker

Step 1: Create Dockerfiles for Each Flask Application

We’ll create three separate Dockerfiles, one for each Flask app.

Flask App 1 (app1.py)

from flask import Flask

app = Flask(__name__)

@app.route("/")
def hello():
    return "Hello from Flask App 1!"

if __name__ == "__main__":
    app.run(host="0.0.0.0", port=5001)

Flask App 2 (app2.py)

from flask import Flask

app = Flask(__name__)

@app.route("/")
def hello():
    return "Hello from Flask App 2!"

if __name__ == "__main__":
    app.run(host="0.0.0.0", port=5002)

Flask App 3 (app3.py)

from flask import Flask

app = Flask(__name__)

@app.route("/")
def hello():
    return "Hello from Flask App 3!"

if __name__ == "__main__":
    app.run(host="0.0.0.0", port=5003)

Each Flask app listens on a different port (5001, 5002, 5003).

Step 2: Dockerfiles for each flask application

Dockerfile for Flask App 1 (Dockerfile.app1)

# Use the official Python image from the Docker Hub
FROM python:3.9-slim

# Set the working directory inside the container
WORKDIR /app

# Copy the current directory contents into the container at /app
COPY app1.py .

# Install Flask inside the container
RUN pip install Flask

# Expose the port the app runs on
EXPOSE 5001

# Run the application
CMD ["python", "app1.py"]

Dockerfile for Flask App 2 (Dockerfile.app2)

FROM python:3.9-slim
WORKDIR /app
COPY app2.py .
RUN pip install Flask
EXPOSE 5002
CMD ["python", "app2.py"]

Dockerfile for Flask App 3 (Dockerfile.app3)

FROM python:3.9-slim
WORKDIR /app
COPY app3.py .
RUN pip install Flask
EXPOSE 5003
CMD ["python", "app3.py"]

Step 3: Create a configuration for HAProxy

global
    log stdout format raw local0
    daemon

defaults
    log     global
    mode    http
    option  httplog
    option  dontlognull
    timeout connect 5000ms
    timeout client  50000ms
    timeout server  50000ms

frontend http_front
    bind *:80
    default_backend servers

backend servers
    balance roundrobin
    server server1 app1:5001 check
    server server2 app2:5002 check
    server server3 app3:5003 check

Explanation:

• frontend http_front: Defines the entry point for incoming traffic. It listens on port 80.
• backend servers: Specifies the servers HAProxy will distribute traffic evenly the three Flask apps (app1, app2, app3). The balance roundrobin directive sets the Round Robin algorithm for load balancing.
• server directives: Lists the backend servers with their IP addresses and ports. The check option allows HAProxy to monitor the health of each server.

Step 4: Create a Dockerfile for HAProxy

Create a Dockerfile for HAProxy (Dockerfile.haproxy)

# Use the official HAProxy image from Docker Hub
FROM haproxy:latest

# Copy the custom HAProxy configuration file into the container
COPY haproxy.cfg /usr/local/etc/haproxy/haproxy.cfg

# Expose the port for HAProxy
EXPOSE 80

Step 5: Create a Dockercompose file

To manage all the containers together, create a docker-compose.yml file

version: '3'

services:
  app1:
    build:
      context: .
      dockerfile: Dockerfile.app1
    container_name: flask_app1
    ports:
      - "5001:5001"

  app2:
    build:
      context: .
      dockerfile: Dockerfile.app2
    container_name: flask_app2
    ports:
      - "5002:5002"

  app3:
    build:
      context: .
      dockerfile: Dockerfile.app3
    container_name: flask_app3
    ports:
      - "5003:5003"

  haproxy:
    build:
      context: .
      dockerfile: Dockerfile.haproxy
    container_name: haproxy
    ports:
      - "80:80"
    depends_on:
      - app1
      - app2
      - app3

Explanation:

• The docker-compose.yml file defines four services: app1, app2, app3, and haproxy.
• Each Flask app is built from its respective Dockerfile and runs on its port.
• HAProxy is configured to wait (depends_on) for all three Flask apps to be up and running.

Step 6: Build and Run the Docker Containers

Run the following commands to build and start all the containers:

# Build and run the containers
docker-compose up --build

This command will build Docker images for all three Flask apps and HAProxy and start them up in the background.

You should see the responses alternating between “Hello from Flask App 1!”, “Hello from Flask App 2!”, and “Hello from Flask App 3!” as HAProxy uses the Round Robin algorithm to distribute requests.

Step 7: Test the Load Balancer

Open your browser or use a tool like curl to make requests to the HAProxy server:

curl http://localhost

Imagine you are managing a busy highway with multiple lanes, and you want to direct specific types of vehicles to particular lanes: trucks to one lane, cars to another, and motorcycles to yet another. In the world of web traffic, this is similar to what Access Control Lists (ACLs) in HAProxy do—they help you direct incoming requests based on specific criteria.

Let’s dive into what ACLs are in HAProxy, why they are essential, and how you can use them effectively with some practical examples.

What are ACLs in HAProxy?

Access Control Lists (ACLs) in HAProxy are rules or conditions that allow you to define patterns to match incoming requests. These rules help you make decisions about how to route or manage traffic within your infrastructure.

Think of ACLs as powerful filters or guards that analyze incoming HTTP requests based on headers, IP addresses, URL paths, or other attributes. By defining ACLs, you can control how requests are handled—for example, sending specific traffic to different backends, applying security rules, or denying access under certain conditions.

Why Use ACLs in HAProxy?

Using ACLs offers several advantages:

1. Granular Control Over Traffic: You can filter and route traffic based on very specific criteria, such as the content of HTTP headers, cookies, or request methods.
2. Security: ACLs can block unwanted traffic, enforce security policies, and prevent malicious access.
3. Performance Optimization: By directing traffic to specific servers optimized for certain types of content, ACLs can help balance the load and improve performance.
4. Flexibility and Scalability: ACLs allow dynamic adaptation to changing traffic patterns or new requirements without significant changes to your infrastructure.

How ACLs Work in HAProxy

ACLs in HAProxy are defined in the configuration file (haproxy.cfg). The syntax is straightforward

acl <name> <criteria>

• <name>: The name you give to your ACL rule, which you will use to reference it in further configuration.
• <criteria>: The condition or match pattern, such as a path, header, method, or IP address.

It either returns True or False.

Examples of ACLs in HAProxy

Let’s look at some practical examples to understand how ACLs work.

Example 1: Routing Traffic Based on URL Path

Suppose you have a web application that serves both static and dynamic content. You want to route all requests for static files (like images, CSS, and JavaScript) to a server optimized for static content, while all other requests should go to a dynamic content server.

Configuration:

frontend http_front
    bind *:80
    acl is_static path_beg /static
    use_backend static_backend if is_static
    default_backend dynamic_backend

backend static_backend
    server static1 127.0.0.1:5001 check

backend dynamic_backend
    server dynamic1 127.0.0.1:5002 check

• ACL Definition: acl is_static path_beg /static : checks if the request URL starts with /static.
• Usage: use_backend static_backend if is_static routes the traffic to the static_backend if the ACL is_static matches. All other requests are routed to the dynamic_backend.

Example 2: Blocking Traffic from Specific IP Addresses

Let’s say you want to block traffic from a range of IP addresses that are known to be malicious.

Configurations

frontend http_front
    bind *:80
    acl block_ip src 192.168.1.0/24
    http-request deny if block_ip
    default_backend web_backend

backend web_backend
    server web1 127.0.0.1:5003 check

ACL Definition:acl block_ip src 192.168.1.0/24 defines an ACL that matches any source IP from the range 192.168.1.0/24.

Usage:http-request deny if block_ip denies the request if it matches the block_ip ACL.

Example 4: Redirecting Traffic Based on Request Method

You might want to redirect all POST requests to a different backend for further processing.

Configurations

frontend http_front
    bind *:80
    acl is_post_method method POST
    use_backend post_backend if is_post_method
    default_backend general_backend

backend post_backend
    server post1 127.0.0.1:5006 check

backend general_backend
    server general1 127.0.0.1:5007 check

Example 5: Redirect Traffic Based on User Agent

Imagine you want to serve a different version of your website to mobile users versus desktop users. You can achieve this by using ACLs that check the User-Agent header in the HTTP request.

Configuration:

frontend http_front
    bind *:80
    acl is_mobile_user_agent req.hdr(User-Agent) -i -m sub Mobile
    use_backend mobile_backend if is_mobile_user_agent
    default_backend desktop_backend

backend mobile_backend
    server mobile1 127.0.0.1:5008 check

backend desktop_backend
    server desktop1 127.0.0.1:5009 check

ACL Definition:acl is_mobile_user_agent req.hdr(User-Agent) -i -m sub Mobile checks if the User-Agent header contains the substring "Mobile" (case-insensitive).

Usage:use_backend mobile_backend if is_mobile_user_agent directs mobile users to mobile_backend and all other users to desktop_backend.

Example 6: Restrict Access to Admin Pages by IP Address

Let’s say you want to allow access to the /admin page only from a specific IP address or range, such as your company’s internal network.

frontend http_front
    bind *:80
    acl is_admin_path path_beg /admin
    acl is_internal_network src 192.168.10.0/24
    http-request deny if is_admin_path !is_internal_network
    default_backend web_backend

backend web_backend
    server web1 127.0.0.1:5015 check

Example with a Flask Application

Let’s see how you can use ACLs with a Flask application to enforce different rules.

Flask Application Setup

You have two Flask apps: app1.py for general requests and app2.py for special requests like form submissions.

app1.py

from flask import Flask

app = Flask(__name__)

@app.route('/')
def index():
    return "Welcome to the main page!"

if __name__ == '__main__':
    app.run(host='0.0.0.0', port=5003)

app2.py:

from flask import Flask

app = Flask(__name__)

@app.route('/submit', methods=['POST'])
def submit_form():
    return "Form submitted successfully!"

if __name__ == '__main__':
    app.run(host='0.0.0.0', port=5004)

HAProxy Configuration with ACLs

frontend http_front
    bind *:80
    acl is_post_method method POST
    acl is_submit_path path_beg /submit
    use_backend post_backend if is_post_method is_submit_path
    default_backend general_backend

backend post_backend
    server app2 127.0.0.1:5004 check

backend general_backend
    server app1 127.0.0.1:5003 check

ACLs:

• is_post_method checks for the POST method.
• is_submit_path checks if the path starts with /submit.

Traffic Handling: The traffic is directed to post_backend if both the ACLs match, otherwise, it goes to general_backend.

Meet Sarah, a backend developer at “BrightApps,” a fast-growing startup specializing in custom web applications. Recently, BrightApps launched a new service called “FitGuru,” a health and fitness platform that quickly gained traction. However, as the platform’s user base started to grow, the team noticed performance issues—page loads were slow, and users began to complain.

Sarah knew that simply scaling up their backend servers might not solve the problem. What they needed was a smarter way to handle incoming traffic and distribute it across their servers. That’s when she decided to dive into the world of Layer 7 (L7) load balancing with HAProxy.

Understanding L7 Load Balancing

Layer 7 load balancing operates at the Application Layer of the OSI model. Unlike Layer 4 (L4) load balancing, which only considers information from the Transport Layer (like IP addresses and ports), an L7 load balancer examines the actual content of the HTTP requests. This deeper inspection allows it to make more intelligent decisions on how to route traffic.

Here’s why Sarah chose L7 load balancing for “FitGuru”:

1. Content-Based Routing: Sarah could route requests to different servers based on the URL path, HTTP headers, cookies, or even specific parameters in the request. For example, requests for video content could be directed to a server optimized for handling media, while API requests could go to a server focused on data processing.
2. SSL Termination: The L7 load balancer could handle the SSL encryption and decryption, relieving the backend servers from this CPU-intensive task.
3. Advanced Health Checks: Sarah could set up health checks that simulate real user traffic to ensure backend servers are actually serving content correctly, not just responding to pings.
4. Enhanced Security: With L7, she could filter out malicious traffic more effectively by inspecting request contents, blocking suspicious patterns, and protecting the app from common web attacks.

Step 1: Sarah’s Plan with HAProxy as an HTTP Proxy

Sarah decided to configure HAProxy as an HTTP proxy. This way, it would operate at Layer 7 and provide advanced traffic management capabilities. She had a few objectives:

• Route traffic based on the URL path to different servers.
• Offload SSL termination to HAProxy.
• Serve static files from specific backend servers and dynamic content from others.

Sarah started with a simple Flask application to test her configuration:

Flask Application Setup

Sarah created two basic Flask apps:

1. Static Content Server (static_app.py):

from flask import Flask, send_from_directory

app = Flask(__name__)

@app.route('/static/<path:filename>')
def serve_static(filename):
    return send_from_directory('static', filename)

if __name__ == '__main__':
    app.run(host='0.0.0.0', port=5001)

This app served static content from a folder named static.

2. Dynamic Content Server (dynamic_app.py):

from flask import Flask

app = Flask(__name__)

@app.route('/')
def home():
    return "Welcome to FitGuru!"

@app.route('/api/data')
def api_data():
    return {"data": "Some dynamic data"}

if __name__ == '__main__':
    app.run(host='0.0.0.0', port=5002)

This app handled dynamic requests like API endpoints and the home page.

Step 2: Configuring HAProxy for HTTP Proxy

Sarah then moved on to configure HAProxy. She created an HAProxy configuration file (haproxy.cfg) to route traffic based on URL paths

global
    log stdout format raw local0
    maxconn 4096

defaults
    mode http
    log global
    option httplog
    option dontlognull
    timeout connect 5000ms
    timeout client  50000ms
    timeout server  50000ms

frontend http_front
    bind *:80

    acl is_static path_beg /static
    use_backend static_backend if is_static
    default_backend dynamic_backend

backend static_backend
    balance roundrobin
    server static1 127.0.0.1:5001 check

backend dynamic_backend
    balance roundrobin
    server dynamic1 127.0.0.1:5002 check

Explanation of the Configuration

1. Frontend Configuration (http_front):
   • The frontend listens on ports 80 (HTTP).
   • An ACL (is_static) is defined to identify requests for static content based on the URL path prefix /static.
   • Requests that match the is_static ACL are routed to the static_backend. All other requests are routed to the dynamic_backend.
2. Backend Configuration:
   • The static_backend handles static content requests and uses a round-robin strategy to distribute traffic between the servers (in this case, just static1).
   • The dynamic_backend handles all other requests in a similar manner.

Step 3: Deploying HAProxy with Docker

Sarah decided to use Docker to deploy HAProxy quickly:

Dockerfile for HAProxy:

FROM haproxy:2.4

COPY haproxy.cfg /usr/local/etc/haproxy/haproxy.cfg

Build and Run:

docker build -t haproxy-http .
docker run -d -p 80:80 -p 443:443 haproxy-http

This command runs HAProxy in a Docker container, listening on ports 80.

Step 4: Testing the Setup

Now, it was time to test!

1. Static Content Test:
   • Sarah visited http://localhost:5000/static/logo.png. The L7 load balancer identified the /static path and routed the request to static_backend.
2. Dynamic Content Test:
   • Visiting http://localhost:5000 or http://localhost:5000/api/data confirmed that requests were routed to the dynamic_backend as expected.

The Result: A Smoother Experience for “FitGuru”

With L7 load balancing in place, “FitGuru” was now more responsive and could efficiently handle the incoming traffic surge:

• Optimized Performance: Static content requests were efficiently served from servers dedicated to that purpose, while dynamic content was processed by more capable machines.
• Enhanced Security: SSL termination was handled by HAProxy, and the backend servers were freed from CPU-intensive encryption tasks.
• Flexible Traffic Management: Sarah could now easily add or modify rules to adapt to changing traffic patterns or requirements.

By implementing Layer 7 load balancing with HAProxy, Sarah provided “FitGuru” with a robust and scalable solution that ensured a seamless user experience, even during peak times. Now, she could confidently tackle the growing demands of their expanding user base, knowing the platform was built to handle whatever traffic came its way.

Layer 7 load balancing was more than just a tool; it was a strategy that allowed Sarah to understand, control, and optimize traffic in a way that best suited their application’s unique needs. And with HAProxy, she had all the flexibility and power she needed to keep “FitGuru” running smoothly.

Meet Jafer, a backend engineer tasked with ensuring the new microservice they are building can handle high traffic smoothly. The microservice is a Flask application that needs to be accessed over TCP, and Jafer decided to use HAProxy to act as a TCP proxy to manage incoming traffic.

This guide will walk you through how Jafer sets up HAProxy to work as a TCP proxy for a sample Flask application.

Why Use HAProxy as a TCP Proxy?

HAProxy as a TCP proxy operates at Layer 4 (Transport Layer) of the OSI model. It forwards raw TCP connections from clients to backend servers without inspecting the contents of the packets. This is ideal for scenarios where:

• You need to handle non-HTTP traffic, such as databases or other TCP-based applications.
• You want to perform load balancing without application-level inspection.
• Your services are using protocols other than HTTP/HTTPS.

In this layer, it can’t read the packets but can identify the ip address of the client.

Step 1: Set Up a Sample Flask Application

First, Jafer created a simple Flask application that listens on a TCP port. Let’s create a file named app.py

from flask import Flask, request

app = Flask(__name__)

@app.route('/', methods=['GET'])
def home():
    return "Hello from Flask over TCP!"

if __name__ == "__main__":
    app.run(host='0.0.0.0', port=5000)  # Run the app on port 5000

Step 2: Dockerize the Flask Application

To make the Flask app easy to deploy, Jafer decided to containerize it using Docker.

Create a Dockerfile

# Use an official Python runtime as a parent image
FROM python:3.9-slim

# Set the working directory
WORKDIR /app

# Copy the current directory contents into the container at /app
COPY . /app

# Install any needed packages specified in requirements.txt
RUN pip install flask

# Make port 5000 available to the world outside this container
EXPOSE 5000

# Run app.py when the container launches
CMD ["python", "app.py"]

To build and run the Docker container, use the following commands

docker build -t flask-app .
docker run -d -p 5000:5000 flask-app

This will start the Flask application on port 5000.

Step 3: Configure HAProxy as a TCP Proxy

Now, Jafer needs to configure HAProxy to act as a TCP proxy for the Flask application.

Create an HAProxy configuration file named haproxy.cfg

global
    log stdout format raw local0
    maxconn 4096

defaults
    mode tcp  # Operating in TCP mode
    log global
    option tcplog
    timeout connect 5000ms
    timeout client  50000ms
    timeout server  50000ms

frontend tcp_front
    bind *:4000  # Bind to port 4000 for incoming TCP traffic
    default_backend flask_backend

backend flask_backend
    balance roundrobin  # Use round-robin load balancing
    server flask1 127.0.0.1:5000 check  # Proxy to Flask app running on port 5000

In this configuration:

• Mode TCP: HAProxy is set to work in TCP mode.
• Frontend: Listens on port 4000 and forwards incoming TCP traffic to the backend.
• Backend: Contains a single server (flask1) where the Flask app is running.

Step 4: Run HAProxy with the Configuration

To start HAProxy with the above configuration, you can use Docker to run HAProxy in a container.

Create a Dockerfile for HAProxy

FROM haproxy:2.4

# Copy the HAProxy configuration file to the container
COPY haproxy.cfg /usr/local/etc/haproxy/haproxy.cfg

Build and run the HAProxy Docker container

docker build -t haproxy-tcp .
docker run -d -p 4000:4000 haproxy-tcp

This will start HAProxy on port 4000, which is configured to proxy TCP traffic to the Flask application running on port 5000.

Step 5: Test the TCP Proxy Setup

To test the setup, open a web browser or use curl to send a request to the HAProxy server

curl http://localhost:4000/

You should see the response

Hello from Flask over TCP!

This confirms that HAProxy is successfully proxying TCP traffic to the Flask application.

Step 6: Scaling Up

If Jafer wants to scale the application to handle more traffic, he can add more backend servers to the haproxy.cfg file

backend flask_backend
    balance roundrobin
    server flask1 127.0.0.1:5000 check
    server flask2 127.0.0.1:5001 check

Jafer could run another instance of the Flask application on a different port (5001), and HAProxy would balance the TCP traffic between the two instances.

Conclusion

By configuring HAProxy as a TCP proxy, Jafer could efficiently manage and balance incoming traffic to their Flask application. This setup ensures scalability and reliability for any TCP-based service, not just HTTP-based ones.

In the world of web applications, imagine you’re running a very popular pizza place. Every evening, customers line up for a delicious slice of pizza. But if your single cashier can’t handle all the orders at once, customers might get frustrated and leave.

What if you could have a system that ensures every customer gets served quickly and efficiently? Enter HAProxy, a tool that helps manage and balance the flow of web traffic so that no single server gets overwhelmed.

Here’s a straightforward guide to understanding HAProxy, installing it, and setting it up to make your web application run smoothly.

What is HAProxy?

HAProxy stands for High Availability Proxy. It’s like a traffic director for your web traffic. It takes incoming requests (like people walking into your pizza place) and decides which server (or pizza station) should handle each request. This way, no single server gets too busy, and everything runs more efficiently.

Why Use HAProxy?

• Handles More Traffic: Distributes incoming traffic across multiple servers so no single one gets overloaded.
• Increases Reliability: If one server fails, HAProxy directs traffic to the remaining servers.
• Improves Performance: Ensures that users get faster responses because the load is spread out.

Installing HAProxy

Here’s how you can install HAProxy on a Linux system:

1. Open a Terminal: You’ll need to access your command line interface to install HAProxy.
2. Install HAProxy: Type the following command and hit enter

sudo apt-get update
sudo apt-get install haproxy

3. Check Installation: Once installed, you can verify that HAProxy is running by typing

sudo systemctl status haproxy

This command shows you the current status of HAProxy, ensuring it’s up and running.

Configuring HAProxy

HAProxy’s configuration file is where you set up how it should handle incoming traffic. This file is usually located at /etc/haproxy/haproxy.cfg. Let’s break down the main parts of this configuration file,

1. The global Section

The global section is like setting the rules for the entire pizza place. It defines general settings for HAProxy itself, such as how it should operate, what kind of logging it should use, and what resources it needs. Here’s an example of what you might see in the global section

global
    log /dev/log local0
    log /dev/log local1 notice
    chroot /var/lib/haproxy
    stats socket /run/haproxy/admin.sock mode 660
    user haproxy
    group haproxy
    daemon

Let’s break it down line by line:

• log /dev/log local0: This line tells HAProxy to send log messages to the system log at /dev/log and to use the local0 logging facility. Logs help you keep track of what’s happening with HAProxy.
• log /dev/log local1 notice: Similar to the previous line, but it uses the local1 logging facility and sets the log level to notice, which is a type of log message indicating important events.
• chroot /var/lib/haproxy: This line tells HAProxy to run in a restricted area of the file system (/var/lib/haproxy). It’s a security measure to limit access to the rest of the system.
• stats socket /run/haproxy/admin.sock mode 660: This sets up a special socket (a kind of communication endpoint) for administrative commands. The mode 660 part defines the permissions for this socket, allowing specific users to manage HAProxy.
• user haproxy: Specifies that HAProxy should run as the user haproxy. Running as a specific user helps with security.
• group haproxy: Similar to the user directive, this specifies that HAProxy should run under the haproxy group.
• daemon: This tells HAProxy to run as a background service, rather than tying up a terminal window.

2. The defaults Section

The defaults section sets up default settings for HAProxy’s operation and is like defining standard procedures for the pizza place. It applies default configurations to both the frontend and backend sections unless overridden. Here’s an example of a defaults section

defaults
    log     global
    option  httplog
    option  dontlognull
    timeout connect 5000ms
    timeout client  50000ms
    timeout server  50000ms

Here’s what each line means:

• log global: Tells HAProxy to use the logging settings defined in the global section for logging.
• option httplog: Enables HTTP-specific logging. This means HAProxy will log details about HTTP requests and responses, which helps with troubleshooting and monitoring.
• option dontlognull: Prevents logging of connections that don’t generate any data (null connections). This keeps the logs cleaner and more relevant.
• timeout connect 5000ms: Sets the maximum time HAProxy will wait when trying to connect to a backend server to 5000 milliseconds (5 seconds). If the connection takes longer, it will be aborted.
• timeout client 50000ms: Defines the maximum time HAProxy will wait for data from the client to 50000 milliseconds (50 seconds). If the client doesn’t send data within this time, the connection will be closed.
• timeout server 50000ms: Similar to timeout client, but it sets the maximum time to wait for data from the server to 50000 milliseconds (50 seconds).

3. Frontend Section

The frontend section defines how HAProxy listens for incoming requests. Think of it as the entrance to your pizza place.

frontend http_front
    bind *:80
    default_backend http_back

• frontend http_front: This is a name for your frontend configuration.
• bind *:80: Tells HAProxy to listen for traffic on port 80 (the standard port for web traffic).
• default_backend http_back: Specifies where the traffic should be sent (to the backend section).

4. Backend Section

The backend section describes where the traffic should be directed. Think of it as the different pizza stations where orders are processed.

backend http_back
    balance roundrobin
    server app1 192.168.1.2:5000 check
    server app2 192.168.1.3:5000 check
    server app3 192.168.1.4:5000 check

• backend http_back: This is a name for your backend configuration.
• balance roundrobin: Distributes traffic evenly across servers.
• server app1 192.168.1.2:5000 check: Specifies a server (app1) at IP address 192.168.1.2 on port 5000. The check option ensures HAProxy checks if the server is healthy before sending traffic to it.
• server app2 and server app3: Additional servers to handle traffic.

Testing Your Configuration

After setting up your configuration, you’ll need to restart HAProxy to apply the changes:

sudo systemctl restart haproxy

To check if everything is working, you can use a web browser or a tool like curl to send requests to HAProxy and see if it correctly distributes them across your servers.