fault Proxy CLI

fault is here to help you uncover and address resiliency issues early in your development cycle. By easily injecting network faults into your application’s daily workflows, fault encourages you to shift resiliency concerns to the left, long before reaching production. The result: more confident engineering teams and applications that gracefully handle the unexpected.

At its core, fault acts as a local proxy you can route traffic through, giving you fine-grained control over conditions like latency, jitter, and faults. Rather than waiting until late-stage testing or worse, customer reports, you can quickly see how your application responds when the network isn’t perfect.

Why fault?

• Speed: Quickly stand up a local test environment with minimal setup.
• Simplicity: Just a few commands let you inject latency or run complex scenarios. No steep learning curve required.
• Extensibility: Tweak parameters, plug into automated tests, and integrate with your existing CI/CD pipelines.
• Insight: Generate structured reports that help pinpoint issues and identify ways to improve resiliency.

Install

Using built binaries

fault is distributed as built bianries on GitHub. Please refer to the documentation.

Using cargo

The fault proxy is installed as follows

cargo +nightly install fault-cli

If you want to enable ebpf support (highly experimental and likely broken):

cargo +nightly install fault-cli --features stealth

In that case you will also need to install the fault ebpf programs.

Run

Once installed, you can start a latency fault inkjection proxy (for instance):

fault run --with-latency --latency-mean 300 --latency-direction ingress --upstream localhost:9090

This is will inject 300ms (mean) latency on network coming back from the host localhost:9090.