Community

Loading...

Azure SignalIR Service

26 Ideas

Auto Scale SignalR Connections and change pricing model based on days to hours

if we want to increase the unit count, only way is to go to portal and do it manually , or we have to use rest api to do the job, please consider this as request, and it will save time to focus on code , less worry on infrastructure part..??If I increased the Unit count to 10 and decrease it in 2 hours, still I�m billed for 10 unit for that day, please change it to hours..??

Increase Free Tier SignalR Connection Limit

We have a medium-sized Azure solution and are using the Azure SignalR free tier in our staging environment just for feature review and testing.We have two app services and a few function apps and immediately exhaust our 20 connection limit before users are able to connect. This prevents us from testing any SignalR features before we deploy to production.We have a paid tier SignalR for production but ideally we'd like to keep this separate from staging and I can't justify a paid tier for staging just yet.It seems like the 20 connection limit is fair but insufficient, please consider increasing the limit. Even a very small increase would satisfy our requirements.

Azure SignalR Service: Balance server connections between instances without causing clients disconnection

The SignalRService documentation indicates that shutting down a service instance (disconnection of server connection) connected to SignalRService automatically disconnects all clients connected virtually to that instance. This behavior prevents the service life cycle from being abstracted from the clients connection life cycle, not being able to restart or redeploy the server service in a transparent way for clients. Let's imagine a cluster with a service with 3 replicas (3 pods) connected to SignalrService and 30 clients connected to it. Ideally those clients will be connected in a balanced way (10 in each pod). Now we need to deploy a new version of the service, without unavailability or disconnection, so the new replicas will gradually rise as the old ones end. Ideally SignalRService should be balancing the connections of the clients that are connected between the instances of the services that it has connected to it, so that it would be able to move the virtual connections of the clients to the new pods and thus remain virtually connected to the new instances without affecting the service. Is this on the signalrService roadmap? Will it be possible at some point? In our team, we consider it the key advantage to abandon our current Redis backbone as a horizontal scaling mechanism for the service and migrate to SignalrService.

Allow Azure SignalR to Connect Only Privately (No Internet, just VNet)

Right now, Azure SignalR is an internet-facing PaaS service. There is no way to remove its exposure to the internet. Allow for Azure SignalR to be able to block all internet traffic and be accessed only via the Azure VNet.

Allow arbitrary number of SignalR standard units to be purchased

Currently Azure customers can only select from a predefined quantity of standard units for the monthly purchase. Those options are 1, 2, 5, 10, 20, 50, ... This is unnecessarily costly to customers when customers have a stable number of concurrent connections is known to be expected month over month, but which doesn't align with the predefined unit quantities. Customers should be able to enter any arbitrary number into the quantity field.

SignalR availbility in Canada Central

Our architecture is in Canada Central, SignalR is available in Canada East which may introduce latency for us. When can we expect it in Canada Central?

Azure SignalR Diagnostics Settings

I would like Azure SignalR to support similar diagnostic settings as other Azure resources. I would like metrics (and logs if available) to be made available for 3rd party integration with other monitoring products (via Event Hubs, Storage Accounts) or with Log Analytics.

This documentation could be improved

This documentation feels like it is assuming prior knowledge/familiarity with your products, which is a problem for me.As a quick example, there is no mention of what a hub is and when it should be used. Is a hub similar to a controller in MVC? can i use a hub to target sub collections of connected users, or do i need multiple hubs for this?

How to set up global distributed environment

We have set up and are using Azure SignalR Service successfully.How can we set up the service globally so that users are redirected to the nearest instance.Can we use Azure Front Door Service to do that? By having the Web App also in another region and it negotiating the client connection to the nearest region?How do the distributed hubs get connected to each other?

Improve the Serverless offering

It would be great to be able to point an Azure SignalR Service to an Azure Function (or any endpoint) for it to invoke when messages are sent from clients. Also including any lifecycle events like when a client connects/disconnects. Otherwise using Azure SignalR Service in a purely serverless way is more effort than it's worth. Currently it seems the best (and arguably the easiest) way to use Azure SignalR Service is by building a .NET Core app to serve as the Hub Server - but in a serverless environment - this isn't feasible.