Use Azure SignalR Service

  • Article

This article shows you how to use SDK in your app server side to connect to SignalR Service when you are using SignalR in your app server.

Create an Azure SignalR Service instance

Follow Quickstart: Use an ARM template to deploy Azure SignalR to create a SignalR service instance.

For ASP.NET Core SignalR

Install the SDK

Run the command to install SignalR Service SDK to your ASP.NET Core project.

dotnet add package Microsoft.Azure.SignalR

In your Startup class, use SignalR Service SDK as the following code snippet.

public void ConfigureServices(IServiceCollection services)
{
    services.AddSignalR()
            .AddAzureSignalR();
}

public void Configure(IApplicationBuilder app)
{
    app.UseEndpoints(routes =>
    {
        routes.MapHub<YourHubClass>("/path_for_your_hub");
    });
}

Configure connection string

There are two approaches to configure SignalR Service's connection string in your application.

  • Set an environment variable with name Azure:SignalR:ConnectionString or Azure__SignalR__ConnectionString.

    • In Azure App Service, put it in application settings.

  • Pass the connection string as a parameter of AddAzureSignalR().

    services.AddSignalR()
        .AddAzureSignalR("<replace with your connection string>");

    or

    services.AddSignalR()
        .AddAzureSignalR(options => options.ConnectionString = "<replace with your connection string>");

Configure options

There are a few options you can customize when using Azure SignalR Service SDK.

ConnectionString

  • Default value is the Azure:SignalR:ConnectionString connectionString or appSetting in web.config file.
  • It can be reconfigured, but make sure the value is NOT hard coded.

InitialHubServerConnectionCount

  • Default value is 5.
  • This option controls the initial count of connections per hub between application server and Azure SignalR Service. Usually keep it as the default value is enough. During runtime, the SDK might start new server connections for performance tuning or load balancing. When you have large number of clients, you can give it a larger number for better throughput. For example, if you have 100,000 clients in total, the connection count can be increased to 10 or 15.

MaxHubServerConnectionCount

  • Default value is null.
  • This option controls the max count of connections allowed per hub between application server and Azure SignalR Service. During runtime, the SDK might start new server connections for performance tuning or load balancing. By default a new server connection starts whenever needed. When the max allowed server connection count is configured, the SDK doesn't start new connections when server connection count reaches the limit.

ApplicationName

  • Default value is null.
  • This option can be useful when you want to share the same Azure SignalR instance for different app servers containing the same hub names. If not set, all the connected app servers are considered to be instances of the same application.

ClaimsProvider

  • Default value is null.
  • This option controls what claims you want to associate with the client connection. It's used when Service SDK generates access token for client in client's negotiate request. By default, all claims from HttpContext.User of the negotiated request are reserved. They can be accessed at Hub.Context.User.
  • Normally you should leave this option as is. Make sure you understand what happens before customizing it.

AccessTokenLifetime

  • Default value is 1 hour.
  • This option controls the valid lifetime of the access token, that Service SDK generates for each client. The access token is returned in the response to client's negotiate request.
  • When ServerSentEvent or LongPolling is used as transport, client connection will be closed due to authentication failure after the expired time. You can increase this value to avoid client disconnect.

AccessTokenAlgorithm

  • Default value is HS256
  • This option provides choice of SecurityAlgorithms when generate access token. Now supported optional values are HS256 and HS512. Note that HS512 is more secure but the generated token is comparatively longer than that using HS256.

ServerStickyMode

  • Default value is Disabled.
  • This option specifies the mode for server sticky. When the client is routed to the server that it first negotiates with, we call it server sticky.
  • In distributed scenarios, there can be multiple app servers connected to one Azure SignalR instance. As internals of client connections explains, client first negotiates with the app server, and then redirects to Azure SignalR to establish the persistent connection. Azure SignalR then finds one app server to serve the client, as Transport Data between client and server explains.
    • When Disabled, the client routes to a random app server. In general, app servers have balanced client connections with this mode. If your scenarios are broadcast or group send, use this default option is enough.
    • When Preferred, Azure SignalR tries to find the app server that the client first negotiates with in a way that no other cost or global routing is needed. This one can be useful when your scenario is sent to connection*. Send to connection can have better performance and lower latency when the sender and the receiver are routed to the same app server.
    • When Required, Azure SignalR always tries to find the app server that the client first negotiates with. This option can be useful when some client context is fetched from negotiate step and stored in memory, and then to be used inside Hubs. However, this option might have performance drawbacks because it requires Azure SignalR to take other efforts to find this particular app server globally, and to keep globally routing traffics between client and server.

GracefulShutdown

GracefulShutdown.Mode
  • Default value is Off
  • This option specifies the behavior after the app server receives a SIGINT (CTRL + C).
  • When set to WaitForClientsClose, instead of stopping the server immediately, we remove it from the Azure SignalR Service to prevent new client connections from being assigned to this server.
  • When set to MigrateClients, in addition, we try migrating client connections to another valid server. The migration will be triggered only after a message is delivered.
    • OnConnected and OnDisconnected are triggered when connections be migrated in/out.
    • IConnectionMigrationFeature can help you identify if the connection is migrated in/out.
    • See our sample codes for detail usage.
GracefulShutdown.Timeout
  • Default value is 30 seconds
  • This option specifies the longest time in waiting for clients to be closed/migrated.

ServiceScaleTimeout

  • Default value is 5 minutes
  • This option specifies the longest time in waiting for dynamic scaling service endpoints, that to affect online clients at minimum. Normally the dynamic scale between single app server and a service endpoint can be finished in seconds, while considering if you have multiple app servers and multiple service endpoints with network jitter and would like to ensure client stability, you can configure this value accordingly.

MaxPollIntervalInSeconds

  • Default value is 5
  • This option defines the max poll interval allowed for LongPolling connections in Azure SignalR Service. If the next poll request doesn't come in within MaxPollIntervalInSeconds, Azure SignalR Service cleans up the client connection.
  • The value is limited to [1, 300].

TransportTypeDetector

  • Default value: All transports are enabled.
  • This option defines a function to customize the transports that clients can use to send HTTP requests.
  • Use this options instead of HttpConnectionDispatcherOptions.Transports to configure transports.

Sample

You can configure above options like the following sample code.

services.AddSignalR()
        .AddAzureSignalR(options =>
            {
                options.InitialHubServerConnectionCount = 10;
                options.AccessTokenLifetime = TimeSpan.FromDays(1);
                options.ClaimsProvider = context => context.User.Claims;

                options.GracefulShutdown.Mode = GracefulShutdownMode.WaitForClientsClose;
                options.GracefulShutdown.Timeout = TimeSpan.FromSeconds(10);
                options.TransportTypeDetector = httpContext => AspNetCore.Http.Connections.HttpTransportType.WebSockets | AspNetCore.Http.Connections.HttpTransportType.LongPolling;
            });

For the legacy ASP.NET SignalR

Note

If it is your first time trying SignalR, we recommend you use the ASP.NET Core SignalR, it is simpler, more reliable, and easier to use.

Install the SDK

Install SignalR Service SDK to your ASP.NET project with Package Manager Console:

Install-Package Microsoft.Azure.SignalR.AspNet

In your Startup class, use SignalR Service SDK as the following code snippet, replace MapSignalR() to MapAzureSignalR({your_applicationName}). Replace {YourApplicationName} to the name of your application, this is the unique name to distinguish this application with your other applications. You can use this.GetType().FullName as the value.

public void Configuration(IAppBuilder app)
{
    app.MapAzureSignalR(this.GetType().FullName);
}

Configure connection string

Set the connection string in the web.config file, to the connectionStrings section:

<configuration>
    <connectionStrings>
        <add name="Azure:SignalR:ConnectionString" connectionString="Endpoint=...;AccessKey=..."/>
    </connectionStrings>
    ...
</configuration>

Configure options

There are a few options you can customize when using Azure SignalR Service SDK.

ConnectionString

  • Default value is the Azure:SignalR:ConnectionString connectionString or appSetting in web.config file.
  • It can be reconfigured, but make sure the value is NOT hard coded.

InitialHubServerConnectionCount

  • Default value is 5.
  • This option controls the initial count of connections per hub between application server and Azure SignalR Service. Usually keep it as the default value is enough. During runtime, the SDK might start new server connections for performance tuning or load balancing. When you have large number of clients, you can give it a larger number for better throughput. For example, if you have 100,000 clients in total, the connection count can be increased to 10 or 15.

MaxHubServerConnectionCount

  • Default value is null.
  • This option controls the max count of connections allowed per hub between application server and Azure SignalR Service. During runtime, the SDK might start new server connections for performance tuning or load balancing. By default a new server connection starts whenever needed. When the max allowed server connection count is configured, the SDK doesn't start new connections when server connection count reaches the limit.

ApplicationName

  • Default value is null.
  • This option can be useful when you want to share the same Azure SignalR instance for different app servers containing the same hub names. If not set, all the connected app servers are considered to be instances of the same application.

ClaimProvider

  • Default value is null.
  • This option controls what claims you want to associate with the client connection. It's used when Service SDK generates access token for client in client's negotiate request. By default, all claims from IOwinContext.Authentication.User of the negotiated request are reserved.
  • Normally you should leave this option as is. Make sure you understand what happens before customizing it.

AccessTokenLifetime

  • Default value is 1 hour.
  • This option controls the valid lifetime of the access token, which Service SDK generates for each client. The access token is returned in the response to client's negotiate request.
  • When ServerSentEvent or LongPolling is used as transport, client connection will be closed due to authentication failure after the expired time. You can increase this value to avoid client disconnect.

AccessTokenAlgorithm

  • Default value is HS256
  • This option provides choice of SecurityAlgorithms when generate access token. Now supported optional values are HS256 and HS512. Note that HS512 is more secure but the generated token is comparatively longer than that using HS256.

ServerStickyMode

  • Default value is Disabled.
  • This option specifies the mode for server sticky. When the client is routed to the server that it first negotiates with, we call it server sticky.
  • In distributed scenarios, there can be multiple app servers connected to one Azure SignalR instance. As internals of client connections explains, client first negotiates with the app server, and then redirects to Azure SignalR to establish the persistent connection. Azure SignalR then finds one app server to serve the client, as Transport Data between client and server explains.
    • When Disabled, the client routes to a random app server. In general, app servers have balanced client connections with this mode. If your scenarios are broadcast or group send, use this default option is enough.
    • When Preferred, Azure SignalR tries to find the app server that the client first negotiates with in a way that no other cost or global routing is needed. This one can be useful when your scenario is sent to connection*. Send to connection can have better performance and lower latency when the sender and the receiver are routed to the same app server.
    • When Required, Azure SignalR always tries to find the app server that the client first negotiates with. This option can be useful when some client context is fetched from negotiate step and stored in memory, and then to be used inside Hubs. However, this option might have performance drawbacks because it requires Azure SignalR to take other efforts to find this particular app server globally, and to keep globally routing traffics between client and server.

MaxPollIntervalInSeconds

  • Default value is 5
  • This option defines the max idle time allowed for inactive connections in Azure SignalR Service. In ASP.NET SignalR, it applies to long polling transport type or reconnection. If the next /reconnect or /poll request doesn't come in within MaxPollIntervalInSeconds, Azure SignalR Service cleans up the client connection.
  • The value is limited to [1, 300].

Sample

You can configure above options like the following sample code.

app.Map("/signalr",subApp => subApp.RunAzureSignalR(this.GetType().FullName, new HubConfiguration(), options =>
{
    options.InitialHubServerConnectionCount = 1;
    options.AccessTokenLifetime = TimeSpan.FromDays(1);
    options.ClaimProvider = context => context.Authentication?.User.Claims;
}));

Scale out application server

With Azure SignalR Service, persistent connections are offloaded from application server so that you can focus on implementing your business logic in hub classes. But you still need to scale out application servers for better performance when handling massive client connections. Below are a few tips for scaling out application servers.

  • Multiple application servers can connect to the same Azure SignalR Service instance.
  • If you'd like to share the same Azure SignalR instance for different applications containing the same hub names, set them with different ApplicationName option. If not set, all the connected app servers are considered to be instances of the same application.
  • As long as the ApplicationName option and the name of the hub class is the same, connections from different application servers are grouped in the same hub.
  • Each client connection is only created in one of the application servers, and messages from that client are only sent to that same application server. If you want to access client information globally (from all application servers), you have to use some centralized storage to save client information from all application servers.

Next steps

In this article, you learn how to use SignalR Service in your applications. Check the following articles to learn more about SignalR Service.

Azure SignalR Service internals

Quickstart: Create a chat room by using SignalR Service