init jenkins input plugin

[ctrlok]

Initial PR. Plugin for getting metrics from Jenkins.
I'm not sure that I did all in right way, so I ask some questions here:

is a good idea to name measurement "jenkins" with a lot of fields like

http.responseCodes.serverError.count
http.responseCodes.serverError.m15_rate
http.responseCodes.serverError.m1_rate
http.responseCodes.serverError.m5_rate
http.responseCodes.serverError.mean_rate

or will be better to create a lot of measurements like http.responseCodes.serverError with fields count, m15_rate, etc?

[ctrlok]

Hm... there is interesting mistake

2017/10/02 11:12:02 http: TLS handshake error from 127.0.0.1:55467: remote error: tls: bad certificate
FAIL	github.com/influxdata/telegraf/plugins/inputs/http_listener	4.449s

[phemmer]

Not going to speak to the measurement name thing, but there are a few minor things I noticed looking over the code.

[phemmer]

err isn't checked.

[phemmer]

This might be a nitpick, though since this gets called a lot it might be significant, but the fmt.Sprintf adds overhead when you can do k1+"."+k2. Same for the similar lines below.

[phemmer]

It looks like there should be a continue here so that we skip the entry with the bad address.

[ctrlok]

Thanks. Actually, I get that line of code from here (https://github.com/influxdata/telegraf/blob/master/plugins/inputs/nginx/nginx.go#L74) so you may want update nginx plugin too.

[danielnelson]

Thanks for the heads up, I updated the nginx plugin b610276

[ctrlok]

@phemmer thanks for your review!

[danielnelson]

It looks like Jenkins uses dropwizard style metrics, these are a little tricky to convert to a style that works well with Telegraf's data model. We have pull requests #2846 and #2369 which both attempt it.

Ideally, we would complete the dropwizard json parser first and use it in this plugin, I was holding it up while I rewrote the configuration loading system but perhaps it is best to finish it up before hand so the plugins can use it.

How about we discuss a general way to translate dropwizard metrics to Telegraf metrics over on #2846 and then we can use it here to parse the Jenkins metrics?

[danielnelson]

@ctrlok: We have merged @atzoum's dropwizard parser on master, we should update this plugin to use it internally for handling the incoming metrics.

[sun-mir]

Any progress on this work?

[russorat]

should we close this in favor of #4022 ?

[danielnelson]

Yes, #4022 will provide a plugin that doesn't require any Jenkins plugins. The Jenkins Metrics Plugin that this PR uses can now be monitored using the http input along with data_format = "dropwizard".