(Quick reminder: this is a public database, so please don't include the expected numbers, or the numbers after it is fixed, in these comments.)

The metrics that you mention are sent by various shell scripts (powerd_suspend and send_metrics_on_resume) using metrics_client, so it's likely that the problem lies there. I'm not aware of any recent changes within powerd that could result in resume-time metrics not being sent -- this code appears to run as part of the same path that's doing everything else.

Number referenced in reporting (2071) is for canary only I think.  Stable is showing much higher count.

Thank you Todd.  Got confused by the (new?) default for the histograms site.