It would be great to be able to provide a filtering expression in the binding of an eventHubTrigger. Using this I could configure my function to only be invoked for certain events that match the filter. For example, I store a property called "eventType" in the application properties of each event. My function is potentially only interested in events of one or more specific types. Currently I code this logic into the function, which works fine. However, it counts as an unnecessary invocation and increases the cost. Ideally the filtering would take place in the host and not incur a function invocation if the event consumed can be filtered out.
The reason for not just using separate event-hubs is the need to preserve ordering for all of related events, hence a single hub for all events pertaining to a certain entity (with the ID of the entity as the PartitionKey).
It would be great to be able to provide a filtering expression in the binding of an eventHubTrigger. Using this I could configure my function to only be invoked for certain events that match the filter. For example, I store a property called "eventType" in the application properties of each event. My function is potentially only interested in events of one or more specific types. Currently I code this logic into the function, which works fine. However, it counts as an unnecessary invocation and increases the cost. Ideally the filtering would take place in the host and not incur a…1 vote
Unsure on this one but marking as unplanned. The filtering would also have to happen at the client to preserve ordering. However even if we did the filtering before the execution it’s still consuming compute so I don’t think it makes sense for it to be “free.” Just because the trigger is doing filtering before execution wouldn’t necessarily negate the price aspect. And since filtering often has some custom parsing / decoding / conditions it’s almost better to do it in code.
I have a function app set up as a private API that requires users to authenticate via Azure AD B2C to be able to make calls. My mobile app does the auth and then provides the Bearer token in the service calls.
This all works but if someone happens to browse an endpoint in a browser I don't want them to be redirected to the login page as the API is only meant to be requested from within the mobile app.
I would like an option to just deny access to the API when a request is not authorised.2 votes
This remains unscheduled, but is still a good item that we would like to get to.
Running R data science functions serverless is key for those organisations which have a deep investment in skills and code in this language. Azure ML Workbench has no R support so the only way to run R in Azure is with Spark Data Bricks or use a VM with R server installed.2 votes
Thanks for your feedback. Currently, we don’t have any plans for R support in Functions. We will monitor the votes on this request to accordingly inform the priority on our backlog.
I am trying to develop an @AzureFunctions using Xamarin Forms. But it is not accepting connections from the cell phone.
How to configure host, port of Azure Functions on Visual Studio 2017 to enable connections from * other than localhost connections?
How to accept connections from local network on Azure Functions (v2) ?
How to configure Host : Port of Azure Functions on Visual Studio 2017 ?
I want that it accepts like ASP.Net Core (.UseUrls("http://+:7071")):
Now listening on: http://[::]:7071
but it is only listening on
Just updating status for accuracy. Still no short term plans to expose outside of localhost. Keep the votes coming.
Could you please add HTTP trigger for on premise (runtime) environments?3 votes
We are currently investing with Kubernetes as our on premises solution with Azure Functions.
Please review docs.microsoft.com/en-us/azure/azure-functi.. and see if this is something that would meet your needs.
If no function is present in a Function App, the “getting started” assistant is shown which allows adding a premade function. However, you can’t specify a name for the resulting function there, which is why you’ll end up with insignificant names such as “HTTPTriggerCSharp1” or “HTTPTriggerJS1”. It would be great if the user could optionally specify a function name in this dialog.2 votes
This makes sense. We’ll add it to the backlog.
Hoping we can get support added for better deployment capabilities with other languages as well as source code management. Great to see the Azure Function tooling finally make it to RTM for C#!2 votes
This remains unplanned.
My service bus will spike at the top of each hour with up to 1000+ messages to get processed immediately. It would be nice for the consumption plan to allow us to scale the # of servers to come online to be based on that queue size. Example: for 1000 messages, I would like 10 servers to come online (1 server per 100 in the queue).2 votes
We don’t currently plan to expose custom scaling logic in the consumption plan. The app service plan does allow you to specify starting scale and can auto-scale from there.
Expose StorageClientFactory or QueueProcessorFactory to enable triggering on client-side-encrypted queues
With WebJobs, a custom StorageClientFactory can set an EncryptionPolicy in the DefaultRequestOptions such that encrypted queue messages are transparently decrypted when triggering a WebJob.
Alternatively, the EncryptionPolicy can be set on a queue by queue basis using a custom QueueProcessorFactory.
If these, or a more general way to modify the JobHostConfiguration were provided, then we would be able to store the storage connection string in Azure KeyVault and use client side encryption with Azure Functions apps.1 vote
Nothing planned but still curious on this one. I think the ask is that you could decrypt on trigger. I expect trigger would work today but you’d have to have decrypt in code per execution. Is hope to decrypt before execution?
- Don't see your idea?