This is likely due to issue 617261 which I started, but didn't yet finish the puppet side of things.

There are also many cloudtail process instances on a machine (e.g. one for each service managed by service_manager). I believe they may use same target fields for metrics (effectively overriding each other's metrics).

Sorry, I created this bug yesterday in a fit of confusion.  I'd actually renamed these metrics months ago (https://codereview.chromium.org/1660453002) and completely forgot about it.  The handful of cloudtail instances in the link above are old ones sending the old metrics.

I fixed the consoles in cl/131682739 and presence/up in https://codereview.chromium.org/2290993002/.

> There are also many cloudtail process instances on a machine

I thought about that already :) http://shortn/_dIHjyMJuEt - the log ID gets munged into the job_name target field.

The following revision refers to this bug:
  https://chromium.googlesource.com/external/github.com/luci/luci-go.git/+/eddd347821b8a2bfc4e94d04681d9d4c24b19fac

commit eddd347821b8a2bfc4e94d04681d9d4c24b19fac
Author: dsansome <dsansome@chromium.org>
Date: Wed Aug 31 01:38:33 2016

Always set presence/up from go tsmon apps.

BUG= 642280 

Review-Url: https://codereview.chromium.org/2290993002

[modify] https://crrev.com/eddd347821b8a2bfc4e94d04681d9d4c24b19fac/appengine/tsmon/middleware.go
[modify] https://crrev.com/eddd347821b8a2bfc4e94d04681d9d4c24b19fac/common/tsmon/metric/metric_test.go
[modify] https://crrev.com/eddd347821b8a2bfc4e94d04681d9d4c24b19fac/common/tsmon/metric/standard_metrics.go
[modify] https://crrev.com/eddd347821b8a2bfc4e94d04681d9d4c24b19fac/common/tsmon/metric/standard_metrics_test.go