UDP Client: Split large points

[joelegasse]

The v2 UDP client will attempt to split points that exceed the configured payload size. It will only do this for points that have a timestamp specified.

This follows from several ancestors: #6913 #7288

Required for all non-trivial PRs

• Rebased/mergable
• Tests pass
• CHANGELOG.md updated

Required only if applicable

You can erase any checkboxes below this note if they are not applicable to your Pull Request.

• InfluxQL Spec updated
• Provide example syntax
• Update man page when modifying a command

[phemmer]

What is UDPBufferSize?

[joelegasse]

A typo that should be updated to UDPPayloadSize :-)

[joelegasse]

My latest commit adds some improvements to the models package to reduce string allocations when writing points.

For now, these are only taken advantage of in the UDP client, but could be used by the HTTP client as well.

I also removed the bytes.Buffer and replaced it with a plain old byte slice.

[joelegasse]

@phemmer @vlasad @maksadbek @sebito91

You've all been involved with this issue for a while now. I think this current change is good enough to be merged. I would really appreciate any testing you could do with your data that would stress the point splitting behavior while ensuring correctness. I've added some changes that should increase performance over previous iterations: mostly memory re-use and in-place scanning, rather than the decode-reencode process that was happening previously.

I still need to update the docs for the client to describe the splitting behavior, along with the reasoning behind the various decisions we've come to over the past 2.5 months :-)

[sebito91]

@joelegasse thank you sir, we'll test this out now and get back to you ASAP.

[jwilder]

@sebito91 Did you get a chance to test this out?

[jwilder]

s/would/would not/?

[sebito91]

Apologies fr the delay, but this works well for us!