CODEWITHSUNDEEP
azureconcurrencycounterazure-functions
Scott MacRitchie
Scott MacRitchie

Reputation: 21

How to build a scalable counter within azure functions

I'm wanting to build a scalable counter that I can increment within Azure functions code. This solution seems like it would do the trick:
http://blog.smarx.com/posts/architecting-scalable-counters-with-windows-azure
However, the above was assuming a classic web server architecture. Any suggestions for how to construct the ID field when using azure functions? Could use the function's invocation id (ExecutionContext.InvocationId) but that would add a counter row for every function call which would get out of hand pretty fast.

Upvotes: 0

Views: 869

Answers (2)

Mikhail Shilkov
Mikhail Shilkov

Reputation: 35154

Unless for purely academic reasons, I would not bother hand-crafting distributed counter in-memory.

Instead, I would use something like Redis and its INCR command. Of course, you can call it from your Azure Function. This will scale to huge number without much effort.

Upvotes: 3

Mike S
Mike S

Reputation: 3169

InvocationId will be per function instance. Azure Functions still run in a process, so you could still get the machine name and process id using traditional means, something like: 

var id = Environment.MachineName + System.Diagnostics.Process.GetCurrentProcess().Id;

The challenge is that with serverless, we specifically try to abstract away from the underlying infrastructure, so instances could potentially run on many different processes over time.

Upvotes: 0

Related Questions