.NET Core web API + Dapper – Retries

Recovering from transient failures with Polly

It’s not uncommon for SQL Azure servers and databases to suffer from “transient failures”. In application logs I have seen these occur during scale up/down events, periods where my application’s performance has been temporarily impacted (but its throughput has not changed), which I assume has been some load balancing going on in the background and when network connectivity has been a bit flakey.

Microsoft has published guidance for building Microservices applications, troubleshooting common AzureSQL errors and improving the resilience of ADO.Net connections which cover different approaches in depth.

For many years I used the Microsoft Enterprise Library Transient Fault Handling Application Block (TOPAZ), then upgraded to the .Net Core Version built by Mo Chavoshi both of which have been retired.

Now I’m using The Polly Project which builds on the concepts of TOPAZ but has been thoroughly re-engineered with lots of extensibility, an active community and modern codebase. Inspired by Ben Hyrman and several other developers I have built a minimalist wrapper for the Dapper Async methods which detects transient errors using the same approach as the Entity Framework Core library.

public static Task<int> ExecuteWithRetryAsync(
			  this IDbConnection connection,
			  string sql,
			  object param = null,
			  IDbTransaction transaction = null,
			  int? commandTimeout = null,
			  CommandType? commandType = null) => RetryPolicy.ExecuteAsync(() => connection.ExecuteAsync(sql, param, transaction, commandTimeout, commandType));

I did think about retry functionality for async methods which returned object/dynamic but have only implemented strongly typed ones for the initial version.

[HttpGet]
public async Task<ActionResult<IAsyncEnumerable<Model.StockItemListDtoV1>>> Get()
{
	IEnumerable<Model.StockItemListDtoV1> response = null;

	try
	{
		using (SqlConnection db = new SqlConnection(this.connectionString))
		{
			response = await db.QueryWithRetryAsync<Model.StockItemListDtoV1>(sql: @"SELECT [StockItemID] as ""ID"", [StockItemName] as ""Name"", [RecommendedRetailPrice], [TaxRate] FROM [Warehouse].[StockItems]", commandType: CommandType.Text);
		}
	}
	catch (SqlException ex)
	{
		logger.LogError(ex, "Retrieving list of StockItems");

		return this.StatusCode(StatusCodes.Status500InternalServerError);
	}

	return this.Ok(response);
}

I have struggled to get reproduceable transient failures without pausing execution in the Visual Studio debugger and tinkering with variables or scaling up/down my databases (limit to how often this can be done) or unplugging the network cable at the wrong time.

The QueueTrigger function retry “rabbit hole”

My first couple of attempts at an Azure Queue Trigger Function which could do retries when an uplink message couldn’t be processed immediately(I didn’t want to throw an exception as this was just a transient issue) didn’t work. I wanted to return the uplink message to the Azure Storage Queue with the initial visibility set to a couple of seconds without throwing an exception.

I tried decorating the method with an Azure Storage Queue output binding but finally settled on the approach below. I can insert a single message into the storage queue and the application would start looping every minute.

public static class UplinkMessageProcessor
{
   const string RunTag = "Processor001";
   static int ConcurrentThreadCount = 0;
   static int MessagesProcessed = 0;

   [FunctionName("UplinkMessageProcessor")]
   public static void Run(
      [QueueTrigger("%UplinkQueueName%", Connection = "AzureStorageConnectionString")] 
      CloudQueueMessage cloudQueueMessage, 
      IBinder binder, ILogger log)
   {
      try
      {
         Interlocked.Increment(ref ConcurrentThreadCount);
         Interlocked.Increment(ref MessagesProcessed);

         log.LogInformation($"{MessagesProcessed} {RunTag} Threads:{ConcurrentThreadCount}");

         CloudQueue outputQueue = binder.Bind<CloudQueue>(new QueueAttribute("%UplinkQueueName%"));

         CloudQueueMessage message = new CloudQueueMessage(cloudQueueMessage.AsString);

         outputQueue.AddMessage(message, initialVisibilityDelay: new TimeSpan(0, 1, 0));
    
         Thread.Sleep(2000);

         Interlocked.Decrement(ref ConcurrentThreadCount);
      }
      catch (Exception ex)
      {
         log.LogError(ex, "Processing of Uplink message failed");

         throw;
      }
   }
}

I used the binder.bind method to get the CloudQueue and CloudQueueMessage details so I could insert a hidden messages back into the queue.

The version of Azure Storage queue libraries used by the function bindings (Sep 2020) may cause some compile time warnings if you select the wrong NuGet package.

Hopefully this has enough keywords that someone trying todo the same thing finds it.

Enterprise Library V6 Data, Exception and Logging with Azure SDK 2.8

I have used the Enterprise library Blocks (which in different forms have been around since 2005) in quite a few projects. Individually the components are pretty good (not always best of breed) but they are well integrated and when used in the way which they were intended to be used work well.

I have just upgraded a client application to Visual Studio 2015 + .Net 4.5 + Enterprise Library V6 and some of the steps were not immediately obvious so hopefully this saves someone else some time. I have sample code for Azure Cloud Service Web and Worker roles.

For both web and worker roles I added the Azure Diagnostics listener to the listener config section of the enterprise library logging settings.

<loggingConfiguration name="" tracingEnabled="true" defaultCategory="General">
	<listeners>
    <add listenerDataType="Microsoft.Practices.EnterpriseLibrary.Logging.Configuration.SystemDiagnosticsTraceListenerData, Microsoft.Practices.EnterpriseLibrary.Logging, Version=6.0.0.0, Culture=neutral, PublicKeyToken=31bf3856ad364e35&amp;amp;quot;
         type="Microsoft.WindowsAzure.Diagnostics.DiagnosticMonitorTraceListener, Microsoft.WindowsAzure.Diagnostics, Version=2.8.0.0, Culture=neutral, PublicKeyToken=31bf3856ad364e35"
         name="AzureDiagnosticTraceListener"/>
   </listeners>
...
</loggingConfiguration>

I then enabled diagnostics on the role and configured the transfer of logs.

Azure Diagnostics configuration dialog

Azure Diagnostic Configuration

This replaces the DiagnosticMonitorConfiguration based approach

DiagnosticMonitorConfiguration diagConfig = DiagnosticMonitor.GetDefaultInitialConfiguration();
diagConfig.Logs.ScheduledTransferLogLevelFilter = LogLevel.Verbose;

// Enable scheduled transfer
diagConfig.Directories.ScheduledTransferPeriod = TimeSpan.FromMinutes(1);
diagConfig.Logs.ScheduledTransferPeriod = TimeSpan.FromMinutes(1);

...
DiagnosticMonitor.Start(&amp;amp;quot;Microsoft.WindowsAzure.Plugins.Diagnostics.ConnectionString&amp;amp;quot;, diagConfig);

For the web role I configured the exception and logging blocks in the Global.asax.cs file


protected void Application_Start()
{
   AreaRegistration.RegisterAllAreas();
   GlobalConfiguration.Configure(WebApiConfig.Register);
   FilterConfig.RegisterGlobalFilters(GlobalFilters.Filters);
   RouteConfig.RegisterRoutes(RouteTable.Routes);
   BundleConfig.RegisterBundles(BundleTable.Bundles);

   // Load the Entlib logging block configuration
   LogWriterFactory logWriterFactory = new LogWriterFactory();
   LogWriter logWriter = logWriterFactory.Create();
   Logger.SetLogWriter(logWriter);

   // Load the Entlib Exception block configuration
   ExceptionPolicyFactory policyFactory = new ExceptionPolicyFactory();
   exManager = policyFactory.CreateManager();
}

For the worker role I configured the exception and logging blocks in the worker role startup

public override bool OnStart()
{
   // Set the maximum number of concurrent connections
   ServicePointManager.DefaultConnectionLimit = 12;
   ...
   LogWriterFactory logWriterFactory = new LogWriterFactory();
   LogWriter logWriter = logWriterFactory.Create();
   Logger.SetLogWriter(logWriter);
   ...
   return result;
}

Then in the webrole webapi2 API controllers you can use embedded SQL or call stored procedures with retries. (This sample code uses the Northwind database and default retry configuration)

public IEnumerable&amp;amp;lt;ProductDto&amp;amp;gt; Get()
{
var products = new List&amp;amp;lt;ProductDto&amp;amp;gt;();

WebApiApplication.exManager.Process(() =&amp;amp;gt;
{
Database db = new DatabaseProviderFactory().Create(&amp;amp;quot;NorthwindInstance&amp;amp;quot;);

RetryPolicy retry = new RetryPolicy&amp;amp;lt;SqlDatabaseTransientErrorDetectionStrategy&amp;amp;gt;(RetryStrategy.DefaultExponential);

var productAccessor = db.CreateSqlStringAccessor(
&amp;amp;quot;SELECT [ProductID],[ProductName],[QuantityPerUnit],[UnitPrice],[UnitsInStock],[Discontinued] FROM Products&amp;amp;quot;,
MapBuilder&amp;amp;lt;ProductDto&amp;amp;gt;
.MapAllProperties()
.Map(p =&amp;amp;gt; p.ID).ToColumn(&amp;amp;quot;ProductID&amp;amp;quot;)
.Map(p =&amp;amp;gt; p.Name).ToColumn(&amp;amp;quot;ProductName&amp;amp;quot;)
.Map(p =&amp;amp;gt; p.QuantityPerUnit).ToColumn(&amp;amp;quot;QuantityPerUnit&amp;amp;quot;)
.Map(p =&amp;amp;gt; p.UnitPrice).ToColumn(&amp;amp;quot;UnitPrice&amp;amp;quot;)
.Map(p =&amp;amp;gt; p.UnitsInStock).ToColumn(&amp;amp;quot;UnitsInStock&amp;amp;quot;)
.Map(p =&amp;amp;gt; p.Discontinued).ToColumn(&amp;amp;quot;Discontinued&amp;amp;quot;)
.Build());
products = retry.ExecuteAction(() =&amp;amp;gt;
{
return productAccessor.Execute().ToList();
});

}, &amp;amp;quot;ProductService&amp;amp;quot;);

return products;
}
public ProductDto Get(int id)
{
ProductDto productDto = null;

WebApiApplication.exManager.Process(() =&amp;amp;gt;
{
Database db = new DatabaseProviderFactory().Create(&amp;amp;quot;NorthwindInstance&amp;amp;quot;);

var productAccessor = db.CreateSqlStringAccessor(
&amp;amp;quot;SELECT [ProductID],[ProductName],[QuantityPerUnit],[UnitPrice],[UnitsInStock],[Discontinued] FROM Products WHERE [ProductID]=@ProductID&amp;amp;quot;,
new ProdductGetByProductIdParameterMapper(db),
MapBuilder&amp;amp;amp;lt;ProductDto&amp;amp;amp;gt;
.MapAllProperties()
.Map(p =&amp;amp;gt; p.ID).ToColumn(&amp;amp;quot;ProductID&amp;amp;quot;)
.Map(p =&amp;amp;gt; p.Name).ToColumn(&amp;amp;quot;ProductName&amp;amp;quot;)
.Map(p =&amp;amp;gt; p.QuantityPerUnit).ToColumn(&amp;amp;quot;QuantityPerUnit&amp;amp;quot;)
.Map(p =&amp;amp;gt; p.UnitPrice).ToColumn(&amp;amp;quot;UnitPrice&amp;amp;quot;)
.Map(p =&amp;amp;gt; p.UnitsInStock).ToColumn(&amp;amp;quot;UnitsInStock&amp;amp;quot;)
.Map(p =&amp;amp;gt; p.Discontinued).ToColumn(&amp;amp;quot;Discontinued&amp;amp;quot;)
.Build());

productDto = productAccessor.Execute(id).SingleOrDefault();

}, &amp;amp;quot;ProductService&amp;amp;quot;);

return productDto;
}