Skip to main content

Polly C# Tutorial


Polly C# Tutorial


Polly is an open-source .Net library used to handle retry logic in your application. There may be cases when you might want to retry a service call if it fails due to transient errors. Transient errors include errors like Server currently busy, database not available, Not enough resource to process requests etc.


Polly provides a callback mechanism where you can retry and configure the number of retries that you want to perform until you record the failure.


How to install Polly Nuget package- 

1. dotnet CLI Command - dotnet add package Polly

2. Install the nuget package from Visual Studio - Search for Polly in 'Manage nuget packages'


Retry Call without using Polly - The below implementation does not look pretty to read as it obscures the actual logic to call the service.


private void CallExternalService()

{

      const int RETRY_ATTEMPTS = 3;

      for (int i =0, i<RETRY_ATTEMPTS,i++)

      {

           Thread.Sleep(i * 100);

           CallExternalService();

           //If Call Success => exit the lopp

           break;

       }

       catch(Exception)

       {

            Log(exception);

       }

}


Examples

Retry Forever - The service is called again and again until it succeeds

Policy.Handle<Exception>()

          .RetryForever()

          .Execute(CallExternalService);


Handle only certain exceptions - 

Policy.Handle<IOException>().Or<UnauthorizedAccessException>()

          .RetryForever()

          .Execute(CallExternalService);


Log the error message during retries - 

Policy.Handle<IOException>().Or<UnauthorizedAccessException>()

          .RetryForever(ex => Log(ex))

          .Execute(CallExternalService);


Retry n Times

Policy.Handle<Exception>()

          .Retry(10)

          .Execute(CallExternalService);


Log the error message during retries 

Policy.Handle<Exception>()

          .Retry(10, (e, i)  => Log($"Error '{e.Message}' at retry #{i}"))

          .Execute(CallExternalService);


WaitandRetry

You can specify the timespan collection stating how long you need to wait before each retry. The retry attempts are based on how many items are present in the collection.

Policy.Handle<Exception>()

          .WaitAndRetry(new []

          .{ TimeSpan.FromMilliseconds(100),  TimeSpan.FromMilliseconds(200) })            .Execute(CallExternalService);


Calculate wait times dynamically - 

Policy.Handle<Exception>()

          .WaitAndRetry(5, count => TimeSpan.FromSeconds(count))

          .Execute(CallExternalService);


Calculate wait times dynamically and an infinite number of retries - 

Policy.Handle<Exception>()

          .WaitAndRetry(count => TimeSpan.FromSeconds(count))

          .Execute(CallExternalService);


Real-world Example - 

var httpTask = Policy

      . HandleResult<(TMessageReturned m, int s)>(t => t.m == null &&                                                   serviceRetryConfig.StatusCodesToRetry.Contains(t.s))

      .WaitAndRetryAsync(serviceRetryConfig.MaxRetry,

          i => TimeSpan.FromMilliseconds(serviceRetryConfig.SlidingIntervalInMs),

          onRetry: (result, timespan, retrycount,context) =>

          {

              count++;

          })

       .ExecuteAsync(() => CallExternalService<TMessageSent, TMessageReturned>(messagebody));              

Labels:

Comments

Post a Comment

Popular posts from this blog

How to clear Visual Studio Cache

How to clear visual studio cache Many times, during development you would face situations where project references are not loaded properly or you get missing/error DLL's. This is because the Component cache gets corrupted randomly and without any warnings. The first option that needs to be done is to clear component cache and restart Visual Studio since the Cache might be holding onto previous DLL versions. Here are the steps on how to clear Visual Studio Cache, Clearing Component Cache: Close all Visual Studio Instances running in your machine. Also, make sure devenv.exe is not running in the Task Manager Delete the Component cache directory - %USERPROFILE%\AppData\Local\Microsoft\VisualStudio\1x.0\ComponentModelCache Restart Visual Studio The above steps should fix the cache issue most of the times, but some times that is not enough and you need to perform the below steps as well. Clearing User's Temp Folder: Open the temp folder in this locatio n -  %USERPROFILE%\AppData\Loc
231 comments
Read more

How to dependency inject to static class

.Net core supports dependency injection. There are many ways that you can inject services like constructor injection, action method injection, property injection. But there will be scenarios where you need to inject dependency services to static classes. For example, injecting services to extension methods. First, create a static class with a one property IServiceProvider type public void ConfigureServices(IServiceCollection services) { services.AddScoped<ILoggerEntry, LoggerEntry>(); services.AddTransient<IMongoRepository, MongoRepository>(); } Second, configure your services in ConfigureServices() method in Startup.cs and define the lifetime of the service instance using either Transient, Scoped or Singleton types. public void ConfigureServices(IServiceCollection services) { services.AddScoped<ILoggerEntry, LoggerEntry>(); services.AddTransient<IMongoRepository, MongoRepository>(); } For the next step to configure the Static class provider proper
Post a Comment
Read more

PCF Health Check Delay - Invocation timeout

In PCF we can set up the health check invocation timeout to execute X seconds after the instance has been created.  But before that, let's understand the PCF Health Check Lifecycle. Stage 1: The application is deployed to PCF Stage 2: When deploying the app, a health check type is specified and optionally a timeout. If a health check type is not specified, then the monitoring process defaults to a port health check Stage 3: Cloud Controller stages, starts, and runs the application Stage 4: Based on the type specified for the app, Cloud Controller configures a health check    that runs periodically for each app instance Stage 5: When Diego starts an app instance, the app health check runs every two seconds until a response indicates that the app instance is healthy or until the health check timeout elapses. The 2-seconds health check interval is not configurable Stage 6: When an app instance becomes healthy, its route is advertised, if applicable.                  Subsequent hea
1 comment
Read more