Traffic management (routing, splitting) in Microservices - Practice Problems & Coding Challenges

Challenge - 5 Problems

🧠 Conceptual

intermediate

Understanding Traffic Routing in Microservices

In a microservices architecture, what is the primary purpose of traffic routing?

A To direct incoming requests to the appropriate microservice based on rules or conditions

B To store user data securely in a database

C To compress data before sending it over the network

D To monitor CPU usage of each microservice

💡 Hint

Think about how requests find the right service to handle them.

Traffic routing ensures that requests are sent to the correct microservice based on defined rules like URL paths, headers, or other conditions.

❓ Architecture

intermediate

Choosing a Traffic Splitting Strategy

You want to deploy a new version of a microservice and gradually shift 20% of traffic to it while keeping 80% on the old version. Which traffic splitting method is best suited for this?

A Use a firewall to block 80% of requests to the new version

B Use weighted routing to split traffic based on percentages

C Use client-side load balancing to randomly pick a version

D Use DNS round-robin to alternate requests between versions

💡 Hint

Consider a method that can precisely control traffic percentages.

Weighted routing allows precise control over what percentage of traffic goes to each version, ideal for gradual rollouts.

❓ scaling

advanced

Scaling Traffic Management for High Load

Your microservices system experiences sudden spikes in traffic. Which design choice helps maintain efficient traffic routing and splitting at scale?

A Route all traffic through the database to decide service versions

B Use a single centralized load balancer for all routing and splitting

C Implement a distributed service mesh with sidecar proxies for routing decisions

D Disable traffic splitting during spikes to reduce complexity

💡 Hint

Think about decentralizing routing logic to handle scale.

A distributed service mesh with sidecar proxies offloads routing decisions to each service instance, improving scalability and fault tolerance.

❓ tradeoff

advanced

Tradeoffs in Traffic Splitting Methods

Which is a key tradeoff when using client-side traffic splitting versus server-side traffic splitting?

A Client-side splitting reduces network latency but increases server load

B Server-side splitting cannot handle weighted traffic percentages

C Server-side splitting requires clients to be updated frequently

D Client-side splitting offers flexibility but can cause inconsistent traffic distribution

💡 Hint

Consider who controls the routing decisions and how that affects consistency.

Client-side splitting lets clients decide where to send requests, which is flexible but can lead to uneven traffic distribution if clients behave differently.

❓ estimation

expert

Estimating Capacity for Traffic Routing Layer

Your microservices system expects 1 million requests per minute. Each routing decision takes 1 millisecond of CPU time. How many CPU cores are needed to handle routing without delay, assuming each core can process 1000 routing decisions per second?

A Approximately 17 cores

B Approximately 10 cores

C Approximately 100 cores

D Approximately 5 cores

💡 Hint

Calculate requests per second and divide by processing capacity per core.

1 million requests per minute = 16,666.67 requests per second. Each core handles 1000 requests per second, so 16.67 cores needed, rounded up to 17.