Closed Bug 808680 Opened 13 years ago Closed 12 years ago

Crash size by product and version

Categories

(Socorro :: Data request, task)

task
Not set
normal

Tracking

(Not tracked)

RESOLVED FIXED

People

(Reporter: laura, Assigned: benjamin)

Details

Attachments

(1 file)

I'd like a report of the crash size by product and version over time. Would like the median, min, and max for each.
We probably should split this by OS as well, as dumps could be vastly different esp. on Windows vs. Mac vs. Linux (not sure if Android matches Linux or not but median/min/max will tell us if we have more variation there).
(In reply to Robert Kaiser (:kairo@mozilla.com) from comment #1) > We probably should split this by OS as well, as dumps could be vastly > different esp. on Windows vs. Mac vs. Linux (not sure if Android matches > Linux or not but median/min/max will tell us if we have more variation > there). OK, I am going to let the current run finish and then will rerun with OS as well. I notice there are "Android_*" fields in the raw data, such as: "Android_Device": "otoro" I see these for both B2G device and FennecAndroid reports.
Status: NEW → ASSIGNED
(In reply to Robert Helmer [:rhelmer] from comment #2) > I notice there are "Android_*" fields in the raw data Yes, we added those to FennecAndroid a long time ago and recently also to B2G - I'm unsure if those are present in B2G desktop builds (which are on other platforms, actually).
I had been running a larger job, but it's timing out on the secondary hbase cluster (talking to metrics folks about it), for the moment here is the last week using the existing pig query (plus B2G). This has MAX raw size but is missing MIN, and I am working on breaking out OS name as well for the next run.
(In reply to Robert Kaiser (:kairo@mozilla.com) from comment #3) > (In reply to Robert Helmer [:rhelmer] from comment #2) > > I notice there are "Android_*" fields in the raw data > > Yes, we added those to FennecAndroid a long time ago and recently also to > B2G - I'm unsure if those are present in B2G desktop builds (which are on > other platforms, actually). I don't see these for B2G Desktop builds - so presumably we can use the presence of these fields to identify Android versus other Linux: https://crash-stats.mozilla.com/report/index/484ece31-7dcc-4b44-9f6b-0ceab2121104
OK "min" is easy to do, "median" is not quite as simple (does not come with pig). However the existing pig job is returning both raw and aggregate data, so I could do a script using that until we figure it out on the pig side. So, next version of the output will have min/max/mean and be broken out by OS as well, while I work on getting larger sample and median working.
Looks like I did a preliminary report but didn't do the followup in comment 6 - sorry! Laura, still needed?
Assignee: rhelmer → nobody
Flags: needinfo?(laura)
I would like to get an updated version of this, yes. It will help with platform planning in 2014.
Flags: needinfo?(laura)
Flags: needinfo?(laura)
Assignee: nobody → benjamin
This is the data from yesterday only. Let me know if you want it broken down or correlated in more detail. Collector script at https://github.com/mozilla/jydoop/blob/master/scripts/dump-sizes.py although I'll probably be moving most of that into socorro-toolbox because I've got a bunch of postprocessing scripts that don't have a good home right now. Firefox: Normal crashes (1518850): Mean: 253611 Median: 146174 Max: 20776263 Plugin hangs (209310): Mean: 335364 Median: 327562 Max: 18637495 Fennec: Normal crashes (71479): Mean: 389419 Median: 369904 Max: 8107920 B2G: Normal crashes (48651): Mean: 208617 Median: 132464 Max: 665320 Thunderbird: Normal crashes (21742): Mean: 217223 Median: 208601 Max: 8921656 SeaMonkey: Normal crashes (1025): Mean: 254963 Median: 163170 Max: 2093744 Plugin hangs (107): Mean: 340209 Median: 330909 Max: 1272171 Camino: Normal crashes (32): Mean: 223998 Median: 239088 Max: 319424 Sunbird: Normal crashes (70): Mean: 102118 Median: 96688 Max: 290748 WeburgMediaCenter: Normal crashes (26): Mean: 234262 Median: 246426 Max: 412823 AFirefox: Normal crashes (49): Mean: 91710 Median: 88620 Max: 143261 Eudora OSE: Normal crashes (26): Mean: 151286 Median: 106252 Max: 1127486 Firefox-17esr: Normal crashes (9): Mean: 292912 Median: 225928 Max: 604256 Saina: Normal crashes (6): Mean: 314800 Median: 301860 Max: 415208 Firefox-Esr: Normal crashes (1): Mean: 791968 Median: 791968 Max: 791968 BAFI: Normal crashes (9): Mean: 67659 Median: 41934 Max: 116980 MetroFirefox: Normal crashes (1): Mean: 481225 Median: 481225 Max: 481225 4399_web: Normal crashes (3): Mean: 111237 Median: 118601 Max: 134334 http;play.pokemonshowdown.com: Normal crashes (2): Mean: 159715 Median: 159715 Max: 160959 Firefox Spaces.ru Edition: Normal crashes (1): Mean: 213108 Median: 213108 Max: 213108 Wyzo: Normal crashes (1): Mean: 50755 Median: 50755 Max: 50755 : Normal crashes (1): Mean: 44537 Median: 44537 Max: 44537
Status: ASSIGNED → RESOLVED
Closed: 12 years ago
Resolution: --- → FIXED
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: