Configure production mode logging with CloudEvents format.

[calavera]

Change the format of the internal logs to follow the CloudEvents specification before serializing them as JSON.

Context

We want to have a consistent format across internal logs and events. We use CloudEvents which is a very simple JSON wrapper with some standard conventions.

What

This changes the format of the logs using the new cloud_events module added in tensorlakeai/tensorlake#390.

Testing

Contribution Checklist

• If a Python package was changed, please run make fmt in the package directory.
• If the server was changed, please run make fmt in server/.
• Make sure all PR Checks are passing.

[eabatalov]

Why do we do this?

[eabatalov]

What's the reason for this change?
Executor logging is independent from tensorlake SDK.
Also we shouldn't add more dependencies on Tensorlake SDK. Executor is an independent service. The only place where it should ideally be using tensorlake sdk is to write integration tests and get FE proto definitions.

[calavera]

What's the reason for this change?

We want to be able to filter all these logs out when customers see the logs in the UI. We need a consistent format for the json payloads that we can identify easily.

Also we shouldn't add more dependencies on Tensorlake SDK.

The code is extremely simple, I can just copy it over. However, Tensorlake is already a dependency here, I'm not adding any more dependencies that we didn't have before.

[calavera]

We've reviewed the logs we send to customers and didn't see these logs there because they're not emmitted by the FE container logs. So we don't need these changes here.