Logic App - Difference between Workflow/Runtime Outbound IP-addresses and Connector Outbound IP-addresses

Question

Can someone please explain to me the difference between the new way that Azure differs between the outbound IP-addresses for the logic apps please.

I think the documentation is so slimmed down in doing a proper explanation of that issue. Reading it it sounds like the IP-addresses have the exact same role in the logic app.

I just need a bit of clarity regarding this point.

So, whats the difference between Workflow/Runtime Outbound IP-addresses and Connector Outbound IP-addresses in a Logic App. Documentation here

Thanks in advance.

Answer 1

Ok, So i have made some progress regarding this question and i think i understand what the difference is. The information that (in my opinion) wasn't clear to me regarding the difference between these two types of IP-addresses in the documentation, was what i'm about to explain in the following parts below.

So, the difference of usage between these two can be explained in how the IP-addresses are used by the logic app in regards to on-platform-resources, and external resources so to speak.

As I've come to understand it as such:

"Connector Outbound Ip-Addresses" - are used by for example, a trigger in your Logic App, that uses an "API-Connector" resource in your resource group or to call a Service Bus queue or something like an FTP. So every connector/trigger/action that needs a "API-Connector" to make a connection/polling or call uses the "Connector Outbound Ip-Addresses".

"Runtime Outbound Ip-Addresses" - are used in whenever the Logic app itself needs to call an external part NOT using a "API-Connector" resource. That would be for example a HTTP-Connector making a REST-call to an external API or system.

Please correct me if i'm wrong on this part and i will correct this post. As for now this is my understanding of the difference in usage between these two.

Answer 2

Connectors today (e.g. SQL, FTP, Salesforce, CRM, whatever) all have static ip addresses per region. However the HTTP and HTTP + Swagger action will send events from a dynamic set of IP Addresses.

We have an item to fix this in the future, but for now best way is to make the call through a connector (including a custom connector) so you can ensure a static IP.