browser process in this: https://chromeperf.appspot.com/report?sid=49182c2739c18ae0d20f5b17cea72719421dba3c1f4643939622ee47359766f1&start_rev=1492542658&end_rev=1495555594

also this: https://chromeperf.appspot.com/report?sid=0671f46c93c8185b4b3af6c2ce74afb523f0858e1a0c398d822180618c27a63b&start_rev=1492542658&end_rev=1495555594

I think the summary charts might have confused you here? When you look at the individual pages, the alerts point to bug 714742. It was WontFixed, looks like there was a device swap.

I don't see any alerts and I cannot bisect the point I intend to bisect. Simon has a fix in flight that he was going to land today.

I saw the device swap bug, but couldn't figure out how to tie back the graphs linked in comment #1 back to it. How would I do that? 

If it's a device swap, then this is no regression.

You need to expand all the charts. The system health dashboard links to the summary. We alert on the individual pages. I think this is why sometimes it looks like there are no alerts.

So, I expanded

ClankInternal/health-plan-clankium-phone/memory.top_10_mobile / memory:chrome:browser_process:reported_by_os:system_memory:proportional_resident_size_avg / background

No alerts.

I expanded 

renderer_process

no alerts.

I expanded the graph that's linked to from Juan's dashboard and I see the alerts you mentioned. That makes sense. If the alerts on all_processes wasn't about changing a device and bisect failed, maybe it would make sense to re-kick using just the x_process metric to see if that helps bisect? Is that the right procedure?