Implement Azure Blob Store

[amankrx]

Description

This PR implements the Azure Blob Store implementation and closely aligns with the AWS S3 implementation.
This utilizes the azure_sdk libraries (azure_core, azure_storage, and azure_storage_blobs). The libraries are still unofficial and being developed, so we should be mindful of any breaking changes in the future.

Apart from that, I started my development by first creating a POC: https://gist.github.com/amankrx/45e7d2a6ed935aa13dda0318681af2ad
This POC tests the get and upload blobs with all the default features disabled for Azure SDK. This creates a custom HttpClient with manually signing the requests to perform the transactions.

Fixes #1542
/claim #1542

Type of change

Please delete options that aren't relevant.

• New feature (non-breaking change which adds functionality)
• This change requires a documentation update

How Has This Been Tested?

I tested it locally using the bazel test command with the azure_blob_backend.json5 file, configured to use the actual blob storage. To test this locally, you should set the environment variable AZURE_STORAGE_KEY as the access key. Additionally, I created a test suite to cover both individual happy paths and error scenarios.

Checklist

• Updated documentation if needed
• Tests added/amended
• Local testing completed
• bazel test //... passes locally
• PR is contained in a single commit, using git amend see some docs

---

This change is 

[CLAassistant]

All committers have signed the CLA.

[laz-001]

Regarding the code, there seems to be quite a lot of duplication going on though. WDYT about something roughly like this?

this would be a rework, ideally doable by a strict refactoring (zero functional change), but this needs to be done before(!) adding new stores, to avoid higher/duplicate effort.

Would you be willing to add a new bounty for this, or do you expect this to happen in context of providing the azure blob store? Personally, I would even split this into 2 bounties: one for the theoretical work (provide the inheritance diagram), another one for the implementation. 3rd and 4ths ones etc. would be for the stores (lower effort).

One thing that is problematic and that you need to know:

@aaronmondal , @MarcusSorealheis, your response-times within the repo and especially your bounties is far too low - be aware that this can alienate contributors (normal ones and bounty-workers).

[amankrx]

The rework for the generalized cloud store is already a WIP. I will raise that PR next week that should separate it out the functionality.

[laz-001]

The rework for the generalized cloud store is already a WIP. I will raise that PR next week that should separate it out the functionality.

It's just that Issues like this one shouldn't take 6 months.

Splitting issues down to chunks and collaborating does the job much faster and with higher quality.

That's all to it.

(and response-times... well...)

[laz-001]

related:

• 💎 Implement Google Cloud Bucket Store #659

@MarcusSorealheis , is this a hickup of algora? the bounty #659 is still listed as 'unpaid' https://algora.io/TraceMachina/bounties

[amankrx]

/build-image nativelink-worker-init

[amankrx]

/build-image

[github-actions]

Image built and pushed!

ghcr.io/TraceMachina/nativelink-worker-init:6f63e7c

[github-actions]

Image built and pushed!

ghcr.io/TraceMachina/nativelink:6f63e7c

[MarcusSorealheis]

Great work!