PollyChaos 1.0.1

PollyChaos

<img src="icon.png" width="100" align="right" />

Chaos engineering for Polly v8. Inject faults and latency into your resilience pipelines to prove your system handles failures gracefully — before production does it for you.

The Simmy-compatible chaos companion for Polly v8. If you used Polly.Contrib.Simmy with Polly v7, this is what you have been waiting for.

Install

dotnet add package PollyChaos

Quick start

using PollyChaos;

// Throw an exception on 10% of calls
var pipeline = new ResiliencePipelineBuilder()
    .AddChaosFault(injectionRate: 0.1)
    .AddRetry(new RetryStrategyOptions { MaxRetryAttempts = 3 })
    .Build();

await pipeline.ExecuteAsync(ct => httpClient.GetAsync("/api/orders", ct), cancellationToken);

Why PollyChaos?

Usage

Fault injection

// Throw ChaosException on 10% of calls
var pipeline = new ResiliencePipelineBuilder()
    .AddChaosFault(injectionRate: 0.1)
    .Build();

// Throw a custom exception
var pipeline = new ResiliencePipelineBuilder()
    .AddChaosFault(injectionRate: 0.05, fault: new HttpRequestException("injected failure"))
    .Build();

Latency injection

// Add a 2-second delay on 5% of calls
var pipeline = new ResiliencePipelineBuilder()
    .AddChaosLatency(injectionRate: 0.05, latency: TimeSpan.FromSeconds(2))
    .Build();

Combined chaos pipeline

var pipeline = new ResiliencePipelineBuilder<HttpResponseMessage>()
    .AddChaosFault<HttpResponseMessage>(injectionRate: 0.05)   // 5% exceptions
    .AddChaosLatency<HttpResponseMessage>(injectionRate: 0.1)  // 10% slow calls
    .AddRetry(new RetryStrategyOptions<HttpResponseMessage> { MaxRetryAttempts = 3 })
    .AddCircuitBreaker(new CircuitBreakerStrategyOptions<HttpResponseMessage>())
    .Build();

Toggle via configuration (feature flag)

Flip chaos on/off without rebuilding the pipeline — ideal for integration test environments:

var chaosOptions = new ChaosFaultStrategyOptions
{
    InjectionRate = 0.2,
    Enabled = config.GetValue<bool>("ChaosEngineering:Enabled"),
    FaultFactory = () => new TimeoutException("chaos: downstream timeout"),
    OnFaultInjected = args =>
    {
        logger.LogWarning("Chaos fault injected for {Operation}", args.Context.OperationKey);
        return ValueTask.CompletedTask;
    },
};

var pipeline = new ResiliencePipelineBuilder()
    .AddChaosFault(chaosOptions)
    .Build();

ASP.NET Core integration

// Program.cs
builder.Services.AddResiliencePipeline("orders-client", (pipelineBuilder, context) =>
{
    var config = context.ServiceProvider.GetRequiredService<IConfiguration>();
    var chaosEnabled = config.GetValue<bool>("ChaosEngineering:Enabled");

    pipelineBuilder
        .AddChaosFault(new ChaosFaultStrategyOptions
        {
            InjectionRate = 0.1,
            Enabled = chaosEnabled,
        })
        .AddChaosLatency(new ChaosLatencyStrategyOptions
        {
            InjectionRate = 0.05,
            Latency = TimeSpan.FromSeconds(3),
            Enabled = chaosEnabled,
        })
        .AddRetry(new RetryStrategyOptions { MaxRetryAttempts = 3 })
        .AddCircuitBreaker(new CircuitBreakerStrategyOptions());
});

Pipeline order

Place chaos strategies outside (before) retry so injected faults are retried — just like real failures:

var pipeline = new ResiliencePipelineBuilder()
    .AddChaosFault(injectionRate: 0.1)   // 1. inject faults
    .AddChaosLatency(injectionRate: 0.1) // 2. inject latency
    .AddRetry(...)                        // 3. retry failures
    .AddCircuitBreaker(...)              // 4. trip if too many failures
    .Build();

Related packages

Support

If PollyChaos helps you ship more resilient software, consider supporting the project:

💼 Need .NET resilience help? Visit solidqualitysolutions.com for consulting and architecture services.

License

MIT