You can use https://sheriff-o-matic.appspot.com/chromium.perf?useMilo (note the useMilo at the end) to see alerts.

The problem has to do with how we generate alerts. The current way of generating them involves downloading master json from a caching service hosted on app engine. chromium.perf's master JSON data is large enough that it sometimes hits the app engine request size limit.

We're working on a new way to generate alerts which uses a different service which gzips the master json data, which allows it to be under the app engine data size limit. I think sean is working on the rollout of this new way to generate alerts, but he got pulled into some work on monorail spam AFAIK.

Should we be defaulting to useMilo for perf?

Not yet. ?useMilo will display *some* alerts but they're not the same ones you'd see without it. 

We have a metric tracking the edit distance between the a-d alerts and the cron alerts (for staging, not live on prod yet) here: https://pcon.corp.google.com/p#chrome-infra/queryplayground?duration=86400&heatmapColorScale=viceroy&legendtable=false&names=Requests%20by%20app%20version&oldHeatmap=false&outputPoints=900&showEditor=true&stacked=true&title=Requests%20by%20app%20version&yAxisLabel=QPM&yAxisMin=0&query=mash&mash=Fetch(Raw('monarch.acquisitions.Task',%20'/chrome/infra/analyzer/cron_alert_diffs'),%20%7B'data_center':%20'appengine',%20'service_name':%20'sheriff-o-matic-staging'%7D)%0A%7C%20Window(Rate('20m'))%0A%7C%20Point(VAL%20*%2060)%0A%7C%20GroupBy(%5B'metric:tree'%5D)

The cron (useMilo) alerts for the chromium.perf tree are still pretty far off from the alerts-dispatcher alerts. Chromium is the closest, and chromium.perf and android are the furthest off. We'll need to dig into that to see what's causing the difference.

The alerting logic is the same for cron and a-d, but cron is fetching the build extract from milo, while a-d is still getting it from CBE.

Got it.  Since this is crucial for perf sheriff workflow, then this really is a P1 and should not go unowned.

Sean is working on spam issues in monorail right now - martiniss@ do you have any extra cycles?

I can look at making the milo alerts correct in SoM. I'm not sure how soon it'll get done though.

Will work on this starting Monday.

martniss@ were you able to start on this monday as planned?  Any updates?

I haven't had a chance to :( I've been pretty busy. 

looking now.

I dug into this somewhat.

It looks like we're running into OOM issues with the cron job, which is killing it? It looks to me like the data is really of out date, but I'm not sure why exactly.

I'll look at this more.

There are workarounds for this. I can run a cron on my machine when the master json  is too large. Lowering priority.

seanmccullough@ actually migrated the chromium.perf tree to run on app engine, so this problem shouldn't happen anymore! Hopefully. Anyways, closing this issue.