New Logz.io output

[idohalevi]

Required for all PRs:

• Signed CLA.
• Associated README.md updated.
• Has appropriate unit tests.

[idohalevi]

@glinton
Any idea why AppVeyor build failed?

[glinton]

Any idea why AppVeyor build failed?

It's tempermental, don't worry about it.

[idohalevi]

Any idea why AppVeyor build failed?

It's tempermental, don't worry about it.

Happy to hear that. Anything else I forgot before it can be reviewed?

[idohalevi]

@glinton Hi, do you know when someone can take a look at this PR?

[Anaisdg]

@danielnelson

[danielnelson]

Thanks for the pull request @idohalevi, we will try to take a look as soon as we can, however we have quite the backlog right now so it may be some time yet.

I notice this plugin pulls in a few new dependencies, some of them seem quite large, in particular leveldb. The logzio-go library also replicates a fair bit of Telegraf functionality, buffering, retries and even stores the buffer to disk. It may make more sense to not use this library and instead write a new client for the logzio HTTP API.

[idohalevi]

@danielnelson
Great feedback. Can you refer me to the retry/buffering code so I can take a look (couldn't find it)?
I'll push again without the logzio-go library

Thank you

[danielnelson]

I only flipped through the code briefly, it is possible I am misunderstanding it, but most of what concerned me was in https://github.com/logzio/logzio-go/blob/master/logziosender.go

Here is the buffering, instead of sending it adds it to another queue: https://github.com/logzio/logzio-go/blob/master/logziosender.go#L193

Retries happen after the dequeue: https://github.com/logzio/logzio-go/blob/master/logziosender.go#L266

[idohalevi]

@danielnelson
I was unclear. I was referring to the telegraf code section (not logz.io) where it handles retries and disk buffering.

[danielnelson]

It is in the agent and internal/models packages.

[idohalevi]

@danielnelson
I removed the logzio-go library and added tests

[glinton]

Thanks

[idohalevi]

@p-zak sorry for the extra commits, it's been a while since I wrote in go. Is it ok now?

[p-zak]

@idohalevi now it is hard to review because of a lot of files has been changed after your rebase

[idohalevi]

@idohalevi now it is hard to review because of a lot of files has been changed after your rebase

@p-zak Not sure how can I change it. Any suggestions? Should we close this one and create a new one with the recent commit only?

[p-zak]

I think that you do not want to change main README.md file. Something went wrong here...

[p-zak]

@idohalevi now it is hard to review because of a lot of files has been changed after your rebase

@p-zak Not sure how can I change it. Any suggestions? Should we close this one and create a new one with the recent commit only?

It seems that sync between your fork and this Telegraf repo has not gone smoothly... I do not know to help with that... @ssoroka do you have any advice?

[ssoroka]

This branch is badly mangled with changes from master and other branches. Could you start a fresh branch off of master and apply only the changes you want to keep?

If you're trying to make the fork off of a local master, you need to update it first. make sure the Telegraf repo is added as a "remote" repo eg git remote add upstream [email protected]:influxdata/telegraf.git, then something like..

git checkout master
git pull --ff-only upstream master

[idohalevi]

@ssoroka @p-zak closing this PR for a cleaner one #8202