All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=844444

(For debugging:) Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?sid=fc00e71c19080f041aa505e97f66d581ee803b5f72f276901319d5ed323ad83d


Bot(s) for this bug's original alert(s):

chromium-rel-mac11-air
chromium-rel-win10
linux-perf
win-high-dpi

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/12f6338c240000

📍 Couldn't reproduce a difference.
https://pinpoint-dot-chromeperf.appspot.com/job/12f6338c240000

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/12a3f922240000

I may have lumped fairly unrelated graphs together. The battor one can probably be ignored but not the memory:chrome ones, running bisect. I suspect they're caused by site isolation.

I think the memory changes are 841252

Well, maybe not.. there's 3 things pointing to this bug it looks like. I was looking at

https://chromeperf.appspot.com/group_report?rev=556326

Which showed me the system_health.memory_desktop	memory:chrome:all_processes:reported_by_os:system_memory:private_footprint_size_avg/browse_news/browse_news_reddit

But there is also https://chromeperf.appspot.com/group_report?bug_id=844444 which shows some gmail-background tests also, which aren't on that revision then.  See what pinpoint says.

I tried to move the first ones over to 841252

Hmm we can remove the battor one and keep those separately in this bug? The revision ranges of the graphs here do not overlap the ones in issue 841252.

Hm actually the reddit thing doesnt look like that bug, it's inverted in my tests locally. Also there are 3 larger improvements at the same time as the regression for that one test. So I don't know what to say about it now.. it looks like noise?

Sorry - I'm not sure how I marked this as WontFix. I think I was only looking at the BattOr one, which I agree can be safely ignored as the ref build moves too.

📍 Found significant differences after each of 5 commits.
https://pinpoint-dot-chromeperf.appspot.com/job/12a3f922240000

Make --site-per-process the default on ToT via fieldtrial_testing_config by lukasza@chromium.org
https://chromium.googlesource.com/chromium/src/+/fb1ccf02ee8ca79e1404abfd3a3a7d540b7d2dbd

predictors: Enable the new preconnect predictor by default by alexilin@chromium.org
https://chromium.googlesource.com/chromium/src/+/5f1039f1025b5294133baab17a578d45c4431063

Disable flaky SystemHealthBenchmarkSmokeTest test by jonross@chromium.org
https://chromium.googlesource.com/chromium/src/+/ddf61b968ad5a0225a9efcf3358dbed0a160a341

Add latency_info to original event list by eirage@chromium.org
https://chromium.googlesource.com/chromium/src/+/4ef6344563acf9961f28f13e36eaebdffd5077e8

Retrieve module information on a background task in ModuleEventSinkImpl by pmonette@chromium.org
https://chromium.googlesource.com/chromium/src/+/9dd3eaada3238a17ed4f1d39ddf76023923b4982

Understanding performance regressions:
  http://g.co/ChromePerformanceRegressions

The pinpoint job in #c13 seems to focus on memory:chrome:all_processes:reported_by_chrome:malloc:allocated_objects_size.  Similarily, the charts linked from #c1 seem to all be focused on memory (rather than something like story:power_avg and/or cpu_time_percentage_avg tracked in  issue 835861 ).

Based on the above, I feel pretty confident in marking this as WontFix - it is expected that site-per-process uses more memory - on average we expect to increase memory usage (e.g. as measured by Memory.Total.PrivateMemoryFootprint) by around 10%.  Please see go/site-isolation-performance#heading=h.vlq4i11mq9q7 (Google-internal, sorry) for more information about impact of Site Isolation on memory metrics.