Max fails and fail timeout in Nginx - Practice Problems & Coding Challenges

Challenge - 5 Problems

🧠 Conceptual

intermediate

Understanding max_fails and fail_timeout in nginx upstream

In nginx, what is the purpose of the max_fails and fail_timeout parameters in an upstream block?

A They set the maximum number of simultaneous connections and the timeout for each connection.

B They specify the maximum number of retries for failed requests and the total timeout for all retries.

C They configure the maximum request size and the timeout for reading client requests.

D They define how many failed attempts to a server are allowed before it is considered down and how long it stays down.

💡 Hint

Think about how nginx decides to stop sending requests to a failing backend server temporarily.

The max_fails parameter sets how many failed attempts to connect to a server are allowed within the fail_timeout period. If the number of failures reaches max_fails, nginx marks the server as down for the duration of fail_timeout.

💻 Command Output

intermediate

Output of nginx error log with max_fails exceeded

Given this nginx upstream configuration:

upstream backend {
    server backend1.example.com max_fails=3 fail_timeout=10s;
}

If backend1 fails 4 times within 10 seconds, what will nginx log?

A 2024/06/01 12:00:00 [info] backend1.example.com is healthy

B 2024/06/01 12:00:00 [warn] backend1.example.com connection refused

C 2024/06/01 12:00:00 [error] upstream server backend1.example.com temporarily disabled while connecting to upstream

D 2024/06/01 12:00:00 [notice] backend1.example.com max_fails reset

💡 Hint

Check what nginx logs when it disables a server due to failures.

When the number of failed attempts exceeds max_fails within fail_timeout, nginx disables the server temporarily and logs an error indicating the server is disabled.

❓ Configuration

advanced

Configuring max_fails and fail_timeout for fast failover

You want nginx to mark a backend server as down after 2 failed attempts and keep it down for 15 seconds. Which configuration snippet achieves this?

A server backend.example.com max_fails=2 fail_timeout=15s;

B server backend.example.com max_fails=5 fail_timeout=2s;

C server backend.example.com max_fails=2 fail_timeout=5s;

D server backend.example.com max_fails=15 fail_timeout=2s;

💡 Hint

Remember max_fails is the failure count, fail_timeout is the duration the server stays down.

Setting max_fails=2 means after 2 failures nginx marks the server down. fail_timeout=15s means the server stays down for 15 seconds before retrying.

❓ Troubleshoot

advanced

Troubleshooting unexpected server down state

An nginx server is marked down too often even though the backend is healthy. The config is:

server backend.example.com max_fails=1 fail_timeout=30s;

What is the most likely cause?

A fail_timeout is too short, so nginx retries too quickly.

B max_fails is too low, causing nginx to mark the server down after a single failure.

C The backend server is actually down, nginx is correct.

D max_fails and fail_timeout are ignored unless proxy_next_upstream is set.

💡 Hint

Consider how sensitive nginx is to failures with these settings.

With max_fails=1, nginx marks the server down after just one failure, which can cause frequent down states even for transient issues.

🔀 Workflow

expert

Sequence of events with max_fails and fail_timeout

Consider this nginx upstream config:

server backend1.example.com max_fails=3 fail_timeout=20s;

Which sequence correctly describes what happens when backend1 fails 3 times in 10 seconds, then recovers after 25 seconds?

A 1,2,4,3

B 1,4,2,3

C 2,1,4,3

D 1,2,3,4

💡 Hint

Think about the order: failure, marking down, timeout expiration, recovery.

First failures happen (1), then nginx disables the server (2). After fail_timeout expires, failure count resets (4), then nginx resumes sending requests (3).