public static string GetEnvShortName(this AbstractApplicationDependencyRegistrar _)
{
return(ASPNETCORE_ENVIRONMENT.ToLower() switch
{
"development" => "dev",
"test" => "test",
"staging" => $"stag",
"production" => $"prod",
_ => throw new NullReferenceException(nameof(ASPNETCORE_ENVIRONMENT))
});
/// <summary> /// default grpc policies /// </summary> /// <param name="registrar"></param> /// <returns></returns> public static List <IAsyncPolicy <HttpResponseMessage> > GenerateDefaultGrpcPolicies(this AbstractApplicationDependencyRegistrar registrar) => GenerateDefaultRefitPolicies(registrar);
public static bool IsDevelopment(this AbstractApplicationDependencyRegistrar _) => ASPNETCORE_ENVIRONMENT.EqualsIgnoreCase("Development");
/// <summary>
/// default rest policies
/// </summary>
/// <returns></returns>
public static List <IAsyncPolicy <HttpResponseMessage> > GenerateDefaultRefitPolicies(this AbstractApplicationDependencyRegistrar _)
{
//隔离策略
//var bulkheadPolicy = Policy.BulkheadAsync<HttpResponseMessage>(10, 100);
//回退策略
//回退也称服务降级，用来指定发生故障时的备用方案。
//目前用不上
//var fallbackPolicy = Policy<string>.Handle<HttpRequestException>().FallbackAsync("substitute data");
//缓存策略
//缓存策略无效
//https://github.com/App-vNext/Polly/wiki/Polly-and-HttpClientFactory?WT.mc_id=-blog-scottha#user-content-use-case-cachep
//var cache = new MemoryCache(Options.Create(new MemoryCacheOptions()));
//var cacheProvider = new MemoryCacheProvider(cache);
//var cachePolicy = Policy.CacheAsync<HttpResponseMessage>(cacheProvider, TimeSpan.FromSeconds(100));
//重试策略,超时或者API返回>500的错误,重试3次。
//重试次数会统计到失败次数
var retryPolicy = Policy.Handle <TimeoutRejectedException>()
.OrResult <HttpResponseMessage>(response => (int)response.StatusCode >= 500)
.WaitAndRetryAsync(new[]
{
TimeSpan.FromSeconds(3),
TimeSpan.FromSeconds(5),
});
//超时策略
var timeoutPolicy = Policy.TimeoutAsync <HttpResponseMessage>(IsDevelopment(_) ? 10 : 9);
//熔断策略
//如下，如果我们的业务代码连续失败50次，就触发熔断(onBreak),就不会再调用我们的业务代码，而是直接抛出BrokenCircuitException异常。
//当熔断时间10分钟后(durationOfBreak)，切换为HalfOpen状态，触发onHalfOpen事件，此时会再调用一次我们的业务代码
//，如果调用成功，则触发onReset事件，并解除熔断，恢复初始状态，否则立即切回熔断状态。
var circuitBreakerPolicy = Policy.Handle <Exception>()
.CircuitBreakerAsync
(
// 熔断前允许出现几次错误
exceptionsAllowedBeforeBreaking: 10
,
// 熔断时间,熔断10分钟
durationOfBreak: TimeSpan.FromMinutes(3)
,
// 熔断时触发
onBreak: (ex, breakDelay) =>
{
//todo
var e = ex;
var delay = breakDelay;
}
,
//熔断恢复时触发
onReset: () =>
{
//todo
}
,
//在熔断时间到了之后触发
onHalfOpen: () =>
{
//todo
}
);
return(new List <IAsyncPolicy <HttpResponseMessage> >()
{
retryPolicy
, timeoutPolicy
, circuitBreakerPolicy.AsAsyncPolicy <HttpResponseMessage>()
});
}
