azure / kubernetes-carbon-intensity-exporter Goto Github PK

Hi there,

I've taken a quick peek at the docker container that is currently here and I'd like to propose a couple of steps to add some security to the image! Especially when it comes to the usage of the username and password.

Are you okay with me picking that up and coming up with a PR in the near future?

kind regards,
Tom

Hi @helayoty @Fei-Guo,
first of all thanks for creating this project! 💚 I've been interested in carbon aware scaling for a while so I was excited to see this project.

I hit an error installing the exporter in kind when using eastus as the region in the helm values.

fail: CarbonAware.DataSources.WattTime.Client.WattTimeClient[0]
      Error getting data from WattTime.  StatusCode: Forbidden. Response: StatusCode: 403, ReasonPhrase: 'Forbidden', Version: 1.1, Content: System.Net.Http.HttpConnectionResponseContent, Headers:
      {
        Date: Fri, 19 May 2023 10:23:29 GMT
        Transfer-Encoding: chunked
        Connection: keep-alive
        Server: nginx
        Content-Type: application/json
      }
fail: CarbonAware.WebApi.Filters.HttpResponseExceptionFilter[0]
      Exception: Error requesting forecast?ba=PJM_ROANOKE

Changing the region to westus prevents the error because it maps to the CAISO_NORTH balancing authority for which WattTime kindly provide data for free.

Is it worth changing the region in the helm values and docs?

Users would still need a WattTime license to access other regions but they can get started with westus without hitting errors. I'm happy to submit a PR if you think this is useful.

Hi there,

I recently heard about this project here: https://rossfairbanks.com/2023/06/05/carbon-aware-temporal-shifting-with-keda/, and first off, amazing initiative, this is exactly what we need, big fan! :)

Now, would it perhaps be an idea to support Electricitymaps next to Wattime as well? They do seem to boast different numbers, but generally offer a more granular and greater coverage of the world in general. Also, in Europe they are known to be quite accurate.

Let me know, curious to hear your answer! Judging from the current code, we could achieve this by adding a little bit of data wrangling before processing it further as is.

kind regards,
Tom