ping @MSantos

@fgiunchedi is it possible to know the exact time that metric was registered? I don't have access to the graphite cluster so I couldn't perform any tests myself.

In T212303#5070934, @MSantos wrote:

@fgiunchedi is it possible to know the exact time that metric was registered? I don't have access to the graphite cluster so I couldn't perform any tests myself.

The metric is sent periodically, e.g. this is from a capture today (the non printable characters at the beginning are the packet headers, can be ignored)

09:04:49.178699 IP maps1002.eqiad.wmnet.38512 > graphite1004.eqiad.wmnet.8125: UDP, length 173
E.....@.@...
@.*
@...p....`.kartotherian.function (metrics) {
    this.metrics = metrics;
    return this;
}:1|c
kartotherian.req.osm-intl.15.png:44|ms
kartotherian.req.osm-intl.11.png.static.1:273|ms

Found potential error and submitted a PR to fix it: https://github.com/kartotherian/kartotherian/pull/109

Change 525841 had a related patch set uploaded (by MSantos; owner: MSantos):
[mediawiki/services/kartotherian@master] Fix statsd metrics

https://gerrit.wikimedia.org/r/525841

Change 525841 merged by MSantos:
[mediawiki/services/kartotherian@master] Fix statsd metrics

https://gerrit.wikimedia.org/r/525841

Thanks @MSantos ! Definitely seeing less of that type of traffic but still some from karthoterian:

U 10.64.16.42:33395 -> 10.64.16.149:8125
kartotherian.req.osm-intl.18.png:198|ms
kartotherian.function(metrics) {
    this.metrics = metrics;
    return this;
}:1|c


U 10.64.16.42:35554 -> 10.64.16.149:8125
kartotherian.req.osm-intl.18.png:119|ms
kartotherian.function(metrics) {
    this.metrics = metrics;
    return this;
}:1|c
kartotherian.req.osm-intl.12.png:39|ms

@fgiunchedi the final fix was deployed a few minutes ago. Please let me know if you run into any issue.