Add traceability_id field to alb log parse

[skbki]

Apparently AWS added new field to ALB log which broke our ingestion pipeline. Only place I found it is AWS Athena guide to parse ALB logs
https://docs.aws.amazon.com/athena/latest/ug/application-load-balancer-logs.html

Closes: #852

[misterek]

This is currently breaking all of our ALB log parsing.

However, is it possible to adjust so that new fields are ignored, not breaking everything? This is apparently being rolled out slowly region to region, so some amount of logs will be broken.

[jszwedko]

Thanks for this contribution @skbki !

I think the traceability_id field needs to be optional since it appears that AWS seems to still be rolling out this change and so there may be users where this field doesn't exist yet (or if they are processing older ALB logs). Your change here makes the field required.

I agree with #862 (comment) that we should also be conservative and just discard unknown fields to future-proof this function, but that could be handled separately from the PR simply adding support for traceability_id.

[skbki]

Good point. I've managed to make the field optional

[jszwedko]

Thanks @skbki . This looks good to me!

[jszwedko]

We'll just need a changelog entry too.

[pront]

Hi @skbki, thank you for this contribution! A cargo fmt should fix the failing CI check.

[jszwedko]

Clippy flagged an issue too (cargo clippy should show it locally).

[skbki]

Corrected issues, should be fine now