Azure Function Python v2

[karynzv]

This is an example of an enrichment process function which consumes events from Event Hub, processes them and insert into CrateDB.

Summary of the changes / Why this is an improvement

Checklist

• Link to issue this PR refers to (if applicable): Fixes #???

[amotl]

Thanks. Didn't do a thorough review, but added one little suggestion. I'd also like to change the folder name and add a little software test case, but this can also happen after merging.

[hammerhead]

Thanks for making it work with the v2 model!

For this to be an example of basic interoperability between CrateDB/Azure Functions, I think we can trim this down quite a bit. The code originates from a specific implementation project that has all sorts of complexity and corner cases. Which I feel we can skip here, trying to make this a concise example showing how easy it is to make the integration work.

Examples:

• Duplicate storage raw/reading: Would it be our default recommendation to store both pre/post enrichment payloads? We could eliminate the raw table, try to insert directly into reading, and if it fails, iterate through the batch row-by-row and insert into error if needed. That would eliminate the whole ValueCache complexity, looking up rows by trace_id and move them from one list to another.
• Speaking of trace_id, that could then also go away if we don't need to correlate rows between raw and reading anymore.
• Data model: I noticed reading doesn't use OBJECT but has the JSON attributes transformed to top-level columns. Using OBJECT(DYNAMIC) is one of the strengths of CrateDB, I suggest using it. That could also reduce the complexity of enrichment if we just pass the original payload through as-is. Or make a very lightweight example of how one can do simple transformations, such as renaming a JSON attribute before insertion, if needed.

[amotl]

Thank you. Two more suggestions from my side, feel free to amend.

• Folder name: topic/serverless/azure-eventhub.
• Squash all commits into a single one.