Make quick summary of build id / OS version / uptimes / etc. available per signature

VERIFIED FIXED in 2.3.3

Status

Socorro
General
P1
normal
VERIFIED FIXED
7 years ago
6 years ago

People

(Reporter: Robert Kaiser, Assigned: brandon)

Tracking

Dependency tree / graph

Firefox Tracking Flags

(Not tracked)

Details

(Whiteboard: Q42011wanted)

Attachments

(5 attachments, 1 obsolete attachment)

(Reporter)

Description

7 years ago
Here are a few comments I got on my blog about getting a quick summary of build id / OS version / uptimes /etc. available on the summary views of signatures:

Jeff Muizelaar:
"It would be nice if it was possible to get more summary information about a crash. For example: What build ids does this crash all occur with? What operating system versions does this all occur with? etc."

Josh Matthews:
"I, like Jeff, would appreciate summaries of the data available - most recent 10 unique build ids, list of unique OS versions, range of uptimes, etc."
(Reporter)

Comment 1

7 years ago
So, this should be on the /report/list page, either in the header or as its own tab, and contain at least, all sorted by highest percentage first (and leaving out zero-entries):

- How many % of those crashes are on what OS+version
  (e.g. 80% Win7, 18% Win Vista, 2% Mac OS X 10.7)
- Ranges of uptimes (e.g. 95% <1min, 3% 1-5min, 1% 15-60min, 1% >1h)
- Product versions (e.g. 50% 5.0, 30% 4.0.1, 10% 6.0a2, 5% 4.0, 3% 7.0a1,
  2% others [in a tooltip: 3.6.17, 3.6.16, 4.0b12, 4.0b8])
- Products (e.g. 99% Firefox, 1% Fennec)
- Architecture (e.g. 90% x64, 10% amd64)
- Process Type (e.g. 100% content)
- Flash versions (e.g. 55% 10.3.181.22, 45% 10.3.181.14)

This are the list that I can think of for now, other info might come up in the discussion, but I think we should get it started with this set and can improve from there later on.

Note that we do have days of build IDs already in the "Table" tab (just badly labelled), so I don't think that's needed in there for the moment.

Marcia, I guess you'd also find this interesting.
What time period would those summaries cover?

Also, note that OS Version and architecture data in the crash reports is not reliable, so I'm not sure how useful that would be.  

I don't think that we even have flash version in Postgres right now ... do we?  Might need to get that from Hadoop.
(Reporter)

Comment 3

7 years ago
(In reply to Josh Berkus from comment #2)
> What time period would those summaries cover?

The same that the whole /report/list being viewed covers as well.

> Also, note that OS Version and architecture data in the crash reports is not
> reliable, so I'm not sure how useful that would be.  

It's what we look at manually now, so it ought to be good enough for what we need.

> I don't think that we even have flash version in Postgres right now ... do
> we?  Might need to get that from Hadoop.

It seems to be in the reports table, because it's in the CSV that AFAIK only does an excerpt from there. And I guess most of the data on /report/list comes from there as well.
(In reply to Robert Kaiser (:kairo@mozilla.com) from comment #3)
> (In reply to Josh Berkus from comment #2)
> > What time period would those summaries cover?
> 
> The same that the whole /report/list being viewed covers as well.
> 
> > Also, note that OS Version and architecture data in the crash reports is not
> > reliable, so I'm not sure how useful that would be.  
> 
> It's what we look at manually now, so it ought to be good enough for what we
> need.
> 
> > I don't think that we even have flash version in Postgres right now ... do
> > we?  Might need to get that from Hadoop.
> 
> It seems to be in the reports table, because it's in the CSV that AFAIK only
> does an excerpt from there. And I guess most of the data on /report/list
> comes from there as well.

Yes reports table has a flash_version column, and /report/list and the CSV do both use this table as well.
Is there anyone who can provide a mapping of Windows version numbers to major versions?  (i.e. 6.5.X --> Windows 7 )?  Right now mostly what I have in the data is the actual under-the-hood version numbers, which are not what the versions of windows are known by.

Anyway, the above will require a new set of matviews to be speedy.  I'm pretty clear on what needs to be in those matviews, but it'll take a couple solid days of work to write them. I have no idea what UI work is required.  As such, I'm assigning this to Laura to determine priority and schedule.
Assignee: nobody → laura
Here's a summary that should be good enough:
http://msdn.microsoft.com/en-us/library/ms724832%28v=vs.85%29.aspx

Where there's overlap (Windows 7 and Windows Server 2008 R2 are both 6.1), I would use the consumer version (Windows 7).

Comment 7

7 years ago
Going to try to pick this up for 2.3.  I'll start by speccing it.
Target Milestone: --- → 2.3

Comment 8

7 years ago
Actions:

- Brandon will write up what we need in the matview and hand off to Josh
- Schalk will put together the view based on what Kairo has asked for above
- Josh will build the matview
- Brandon will write the gluecode between the matview and the UI
Can you provide more samples for Process Type (e.g. 100% content)
(Reporter)

Comment 10

7 years ago
process_type is a field we get in crash reports directly, with the values right now (and for the foreseeable future) being either '' (meaning 'browser' for us), 'plugin', or 'content'.
This tells us is that is the main "browser" (or application) process, or if it's a plugin or a content (website) process.
Created attachment 563820 [details]
First prototype

Please see the screen shot attached. This is a first prototype. The bars can/will have tooltips when hovered over with more details as needed.

Let me know your thoughts, we can also do this as plain tables if that is more preferred. Looking forward to hearing from you.

Updated

7 years ago
Assignee: laura → sneethling
Target Milestone: 2.3 → 2.3.1
(Reporter)

Comment 12

7 years ago
(In reply to Schalk Neethling from comment #11)
> Let me know your thoughts, we can also do this as plain tables if that is
> more preferred. Looking forward to hearing from you.

I think plain tables would be a good start - and I'd like to have all the data sets from comment #1 should be on there, if possible. Having graphs is interesting, though. Given that it's percentages, pie charts could potentially be even more helpful, I guess.
[:kairo@mozilla.com] We can have both ;) We have the table format as the default but then, offer the user to switch between tabular and pie chart view on demand.

I am thinking this data is not realtime, but even if it is, the info for either is a simple ajax call away

Started 'spec'ing it here: https://wiki.mozilla.org/Socorro/SocorroSignatureDashboard

Feel free too edit, add and delete
Created attachment 565905 [details]
Second prototype
Added a screen shot where the data is displayed using tables. Thougts? Do you think this needs to be one big table instead of the individual one's? Am I displaying all the data that is required?

Wiki page is here: https://wiki.mozilla.org/Socorro/SocorroSignatureDashboard
(Reporter)

Comment 16

7 years ago
Hmm, with the spacing that this second prototype has, I'd think we could even add one more column that has the actual number.

The headers including the "days" selector are part of the overall page already, so I'm just ignoring those right now when looking at this. :)

From the general view, this looks good, I'd think though that the table footers are not necessary as they only repeat info that's already there.

And another tiny thing: I'd like the uptime ranges to be <1min, 1-5min, 5-15min, 15-60min, >1h
The categories between 15 and 60 min don't give us too much extra info in the end.
Created attachment 565918 [details]
Second prototype revised

Removed footers from tables
Changed ranges to reflect those suggested by [:kairo]
Just following on from IRC: When you say "the actual number" - Do you mean the number of reported crashes that make up the percentage?
Created attachment 565919 [details]
Second prototype revision 2

Added new 'Number of Crashes' column
Attachment #565918 - Attachment is obsolete: true
(Reporter)

Comment 20

7 years ago
(In reply to Schalk Neethling from comment #19)
> Created attachment 565919 [details]
> Second prototype revision 2

I like this one!
Asked jberkus to comment on the best way to build the data shown in the mock.
I currently use the JSON data format as below:

Hope the formatting keeps ;)

{
	"percentageByOs" : [
		{
			"os" : "Windows 7",
			"percentage" : "50",
			"numberOfCrashes" : "1000"
		}, 
		{
			"os" : "Windows Vista",
			"percentage" : "25",
			"numberOfCrashes" : "500"
		},
		{
			"os" : "OSX",
			"percentage" : "25",
			"numberOfCrashes" : "500"
		}
	],
	"uptimeRange" : [
		{
			"range" : "< 1 min",
			"percentage" : "20",
			"numberOfCrashes" : "1000"
		},
		{
			"range" : "1 - 5min",
			"percentage" : "20",
			"numberOfCrashes" : "1000"
		},
		{
			"range" : "5 - 15min",
			"percentage" : "20",
			"numberOfCrashes" : "1000"
		},
		{
			"range" : "15 - 60min",
			"percentage" : "20",
			"numberOfCrashes" : "1000"
		},
		{
			"range" : "> 1 hour",
			"percentage" : "20",
			"numberOfCrashes" : "1000"
		}
	],
	"productVersions" : [
		{
			"product" : "Firefox",
			"version" : "3.6",
			"percentage" : "50",
			"numberOfCrashes" : "1000"
		},
		{
			"product" : "Firefox",
			"version" : "7.0",
			"percentage" : "25",
			"numberOfCrashes" : "500"
		},
		{
			"product" : "Fennec",
			"version" : "4.1",
			"percentage" : "25",
			"numberOfCrashes" : "500"
		}
	]
}
That's 3 separate matviews.  Other than needing to write the generator code, it's fairly simple and direct.  I'll want to bundle this in with re-plumbing the matview generation, but it's not really a technical challenge.

Populating the view itself will require a windowing query over the matview, so I'll need to help Schalk with how to do that.

Missing Data: The only piece of data we're missing to populate that matview is being able to match the real OS versions with user-friendly names.  For example, what we have in the database is NOT "Windows Vista", it's "Windows v5.51".  I don't have the knowledge to match up the real versions to the marketing versions; does anyone else? (note that this is primarily a Windows problem).

Some questions on what data we want in the matview:

1) At what % do we want to cut off products and OS versions and not display them?  i.e. how far do we want to go down into 0.003% before we ignore it?

2) Do we want specific versions of OSX and Linux as well as Windows?  We have that data.

3) What OS names do we want to use for Fennec?  Right now OS for Fennec is largely ignored.  OSes for Fennec have completely different names from Firefox.
Oh, one more:

4) For product, do we want to list just major version (7.0) or do we want to list by release? (7.0.1, 7.0b4)
Progressing, but won't land for 2.3.1.
Target Milestone: 2.3.1 → 2.3.2
[:laura] What are the next steps with regards to this?
Assigning over to Brandon for implementation.
Assignee: sneethling → bsavage
[:laura] Please let me know what to do with the prototype code or, would Brandon be writing the code that will eventually provide me with the JSON I use to build the tables on the client?
(Reporter)

Comment 29

6 years ago
(In reply to Josh Berkus from comment #23)
> Missing Data: The only piece of data we're missing to populate that matview
> is being able to match the real OS versions with user-friendly names.  For
> example, what we have in the database is NOT "Windows Vista", it's "Windows
> v5.51".  I don't have the knowledge to match up the real versions to the
> marketing versions; does anyone else? (note that this is primarily a Windows
> problem).

We probably should do a static table to do that matching and have it manually updated when needed, it shouldn't change very often.

> 1) At what % do we want to cut off products and OS versions and not display
> them?  i.e. how far do we want to go down into 0.003% before we ignore it?

I guess we can cut off at 1% when we display "others <1%" or so where others do exist.

> 2) Do we want specific versions of OSX and Linux as well as Windows?  We
> have that data.

We'd at least want the "major" versions of OSX, which happens to be what some might call "minor versions", i.e. 10.5, 10.6, 10.7 - for Linux, I think it doesn't make sense to differentiate by kernel versions, and that's actually what we do get.

> 3) What OS names do we want to use for Fennec?  Right now OS for Fennec is
> largely ignored.  OSes for Fennec have completely different names from
> Firefox.

Right now, don't treat them differently, even if that makes us list "100% Linux" for now. We might want to later go and regex-match the OS version string and replace with something we know, but right now let's don't go that far (maybe think about being able to reuse the same code we are using for the Windows matching, though).

> 4) For product, do we want to list just major version (7.0) or do we want to
> list by release? (7.0.1, 7.0b4)

Let's go with the "full" version, i.e. the latter of what you are mentioning.

Updated

6 years ago
Whiteboard: Q42011wanted

Updated

6 years ago
Priority: -- → P1
I should be able to do the matviews early next week.
(Assignee)

Comment 31

6 years ago
Per an IRC conversation with Josh, we're waiting on the matviews for some higher priority issues. This is unlikely to make 2.3.2 but is possible for 2.3.3; Josh indicated that he'd be able to start working on the code later in the week.
(In reply to Robert Kaiser (:kairo@mozilla.com) from comment #29)
> (In reply to Josh Berkus from comment #23)
> > Missing Data: The only piece of data we're missing to populate that matview
> > is being able to match the real OS versions with user-friendly names.  For
> > example, what we have in the database is NOT "Windows Vista", it's "Windows
> > v5.51".  I don't have the knowledge to match up the real versions to the
> > marketing versions; does anyone else? (note that this is primarily a Windows
> > problem).
> 
> We probably should do a static table to do that matching and have it
> manually updated when needed, it shouldn't change very often.

I still need someone to provide me with version numbers and names.  I do not have this knowledge nor do I know where to find it, and casual Googling has turned up nothing.  It's fine to have a static table, but someone needs to populate it.
StephenD just supplied this.  Unless told otherwise, I'll be using the verision numbers listed here:

http://www.msigeek.com/442/windows-os-version-numbers
(Reporter)

Comment 34

6 years ago
(In reply to Josh Berkus from comment #33)
> StephenD just supplied this.  Unless told otherwise, I'll be using the
> verision numbers listed here:
> 
> http://www.msigeek.com/442/windows-os-version-numbers

Let's try it - and make all that are not fitting with that use the pure version number. We should be able to augment the list as needed.
(Reporter)

Comment 35

6 years ago
Oh, and we only support Windows 200 and up for Firefox anyhow, so you don't need the versions below 5.
(Reporter)

Comment 36

6 years ago
And http://www.gaijin.at/en/lstwinver.php knows even more versions, I think we could just go with this though:

Windows 2000          5.0.*
Windows XP            5.1.*
Windows Server 2003   5.2.*
Windows Vista         6.0.*
Windows 7             6.1.*

Windows 8 is rumored to be 6.2.*, but I don't think we can know yet for sure.

Updated

6 years ago
Assignee: bsavage → josh
Whiteboard: Q42011wanted → Q42011wanted [2.3.3]
Target Milestone: 2.3.2 → 2.4
This is waiting on my matviews

Updated

6 years ago
Whiteboard: Q42011wanted [2.3.3] → Q42011wanted
Target Milestone: 2.4 → 2.3.3
Brandon,

The following matviews are now available for testing on DevDB:

os_signature_counts
uptime_signature_counts
product_signature_counts

I will work on an example query for how to use these.
Brandon,

Ok, these look really really hairy, but they're very fast to execute, and produce display-ready output.   Here's three example queries to use the 3 matviews.  Note that these assume that you have a signature_id to work with; if you need to use the actual signature string, let me know and I'll provide modified versions.


WITH counts AS (
	SELECT os_version_string, report_count,
		sum(report_count) over () as total_count
	FROM os_signature_counts
	WHERE signature_id = 1239904
		AND report_date BETWEEN '2011-11-03' and '2011-11-06'
)
SELECT os_version_string, 
	round(sum(report_count)*100/max(total_count)::numeric, 1) as report_percent
FROM counts
GROUP BY os_version_string
HAVING sum(report_count)*100/max(total_count)::numeric >= 1.0
ORDER BY report_percent DESC;


WITH counts AS (
	SELECT uptime_level, report_count,
		sum(report_count) over () as total_count
	FROM uptime_signature_counts
	WHERE signature_id = 1239904
		AND report_date BETWEEN '2011-11-03' and '2011-11-06'
)
SELECT uptime_string, 
	round(sum(report_count)*100/max(total_count)::numeric, 1) as report_percent
FROM counts
	JOIN uptime_levels USING (uptime_level)
GROUP BY uptime_string
HAVING sum(report_count)*100/max(total_count)::numeric >= 1.0
ORDER BY report_percent DESC;


WITH counts AS (
	SELECT product_version_id, report_count,
		sum(report_count) over () as total_count
	FROM product_signature_counts
	WHERE signature_id = 1239904
		AND report_date BETWEEN '2011-11-03' and '2011-11-06'
)
SELECT product_name || ' ' || version_string as product_full_name, 
	round(sum(report_count)*100/max(total_count)::numeric, 1) as report_percent
FROM counts
	JOIN product_versions USING (product_version_id)
GROUP BY product_name, version_string
HAVING sum(report_count)*100/max(total_count)::numeric >= 1.0
ORDER BY report_percent DESC;

Updated

6 years ago
Assignee: josh → bsavage
[:brandon] I can supply the front end code, just tell me where to send it. Thanks to rhelmer's docs and adding new reports to the front end, this process will be much smoother going forward.
Adjusted queries per Brandon's request:

WITH counts AS (
	SELECT os_version_string, report_count,
		sum(report_count) over () as total_count
	FROM os_signature_counts
	WHERE signature_id = 1239904
		AND report_date BETWEEN '2011-11-03' and '2011-11-06'
)
SELECT os_version_string, 
	round(sum(report_count)*100/max(total_count)::numeric, 1) as report_percent,
	sum(report_count) as report_count
FROM counts
GROUP BY os_version_string
HAVING sum(report_count)*100/max(total_count)::numeric >= 1.0
ORDER BY report_count DESC;


WITH counts AS (
	SELECT uptime_level, report_count,
		sum(report_count) over () as total_count
	FROM uptime_signature_counts
	WHERE signature_id = 1239904
		AND report_date BETWEEN '2011-11-03' and '2011-11-06'
)
SELECT uptime_string, 
	round(sum(report_count)*100/max(total_count)::numeric, 1) as report_percent,
	sum(report_count) as report_count
FROM counts
	JOIN uptime_levels USING (uptime_level)
GROUP BY uptime_string
HAVING sum(report_count)*100/max(total_count)::numeric >= 1.0
ORDER BY report_count DESC;


WITH counts AS (
	SELECT product_version_id, report_count,
		sum(report_count) over () as total_count
	FROM product_signature_counts
	WHERE signature_id = 1239904
		AND report_date BETWEEN '2011-11-03' and '2011-11-06'
)
SELECT product_name || ' ' || version_string as product_full_name, 
	round(sum(report_count)*100/max(total_count)::numeric, 1) as report_percent,
	sum(report_count) as report_count
FROM counts
	JOIN product_versions USING (product_version_id)
GROUP BY product_name, version_string
HAVING sum(report_count)*100/max(total_count)::numeric >= 1.0
ORDER BY report_percent DESC;
(Assignee)

Comment 43

6 years ago
The code and layout have been installed, tested and committed to my branch https://github.com/brandonsavage/socorro/commit/9a0fb69ea23d979c90f42d83e281ad4cb403942e. There are a few UI issues to be worked out, which I'll need to work on with espressive, who I believe has gone home for the night. I've requested code review from :rhelmer to make sure this can make it in before freeze.
(In reply to Brandon Savage [:brandon] from comment #43)
> The code and layout have been installed, tested and committed to my branch
> https://github.com/brandonsavage/socorro/commit/
> 9a0fb69ea23d979c90f42d83e281ad4cb403942e. There are a few UI issues to be
> worked out, which I'll need to work on with espressive, who I believe has
> gone home for the night. I've requested code review from :rhelmer to make
> sure this can make it in before freeze.

Commented on github, the gist is that we should add a middleware service to do the SQL and return JSON (rather than having the PHP app connect to PostgreSQL directly and do the work to turn it into JSON), since we're trying to get the frontend to use the middleware instead of talking directly to postgres/hbase/etc.

I could help whip up a service for this if we don't have an existing one, let me know.
[:brandon] [:rhelmer] How would I one, be able to get the items from the commit to my dev instances and two, get the changes in Brandon's branch into mine locally, simply clone his fork? Thanks
(In reply to Robert Helmer [:rhelmer] from comment #44)
> (In reply to Brandon Savage [:brandon] from comment #43)
> > The code and layout have been installed, tested and committed to my branch
> > https://github.com/brandonsavage/socorro/commit/
> > 9a0fb69ea23d979c90f42d83e281ad4cb403942e. There are a few UI issues to be
> > worked out, which I'll need to work on with espressive, who I believe has
> > gone home for the night. I've requested code review from :rhelmer to make
> > sure this can make it in before freeze.
> 
> Commented on github, the gist is that we should add a middleware service to
> do the SQL and return JSON (rather than having the PHP app connect to
> PostgreSQL directly and do the work to turn it into JSON), since we're
> trying to get the frontend to use the middleware instead of talking directly
> to postgres/hbase/etc.
> 
> I could help whip up a service for this if we don't have an existing one,
> let me know.

OK per IRC, I am willing to take this to avoid conflicting with the mware refactoring going on, and we will finally move everything out in bug 701600.
(In reply to Schalk Neethling from comment #45)
> [:brandon] [:rhelmer] How would I one, be able to get the items from the
> commit to my dev instances and two, get the changes in Brandon's branch into
> mine locally, simply clone his fork? Thanks

Yes, or you could pull from his branch, or you can get a diff of:
https://github.com/mozilla/socorro/pull/145

By appending ".diff" to the URL:
https://github.com/mozilla/socorro/pull/145.diff

Then you could do "curl https://github.com/mozilla/socorro/pull/145.diff | patch -p1" to apply it.

Comment 48

6 years ago
Commits pushed to https://github.com/mozilla/socorro

https://github.com/mozilla/socorro/commit/93592404724849549dda8c24ecce0a53355bd979
Bug 642336 - Adding report to Socorro for a quick summary for a signature.

https://github.com/mozilla/socorro/commit/e05c57c7fb063861ffdfbe32c7dcfe22203939f1
Merge pull request #145 from brandonsavage/bug642336

Bug 642336 - Adding report to Socorro for a quick summary for a signature
(Assignee)

Updated

6 years ago
Status: NEW → RESOLVED
Last Resolved: 6 years ago
Resolution: --- → FIXED
Verified on dev and stage.
Status: RESOLVED → VERIFIED
(Reporter)

Comment 50

6 years ago
One thing that looks strange with this is that the data used in this summary is different from the data used for the other tabs on the same page (and actually the original request was to use the same data).

See e.g. https://crash-stats.allizom.org/report/list?range_value=7&range_unit=days&date=2011-11-08&signature=arena_dalloc_small%20%7C%20arena_dalloc%20%7C%20free%20%7C%20nsPrefBranch%3A%3A%60vector%20deleting%20destructor%27%27%28unsigned%20int%29&version=Firefox%3A8.0 talking about 133 crashes across multiple versions, when the link leading to it from the Firefox 8.0 topcrash report has 23 crashes and the "Reports" table also shows only those 23.
Kairo,

That's because the Signature Summary shows *all* products, not just FF 8.0.  Otherwise the "summary by product" wouldn't make much sense.
(Reporter)

Comment 52

6 years ago
Well, this is not what I originally have requested, then.

On https://crash-stats.allizom.org/report/list?signature=arena_dalloc_small%20%7C%20arena_dalloc%20%7C%20free%20%7C%20nsPrefBranch%3A%3A%60vector%20deleting%20destructor%27%27%28unsigned%20int%29 the products split makes sense, on https://crash-stats.allizom.org/report/list?range_value=7&range_unit=days&date=2011-11-08&signature=arena_dalloc_small%20%7C%20arena_dalloc%20%7C%20free%20%7C%20nsPrefBranch%3A%3A%60vector%20deleting%20destructor%27%27%28unsigned%20int%29&version=Firefox%3A8.0 it of course doesn't too much, but it doesn't matter if it reflects correctly what is in the set.

I'm also under the impression that the latter query actually doesn't include the 23 crashes listed in the "Reports" tab, but maybe that's some stage wonkiness.
Have discussed the above with Kairo.  There are some other data issues to hash out; but right now there's a more significant UI bug:

If you click on a different number of weeks, the 3-table summary does not refresh: it shows one week regardless of the number of weeks selected.
Status: VERIFIED → REOPENED
Resolution: FIXED → ---
Reverting to verified because the miscount issue already has its own bug, 702359
Status: REOPENED → RESOLVED
Last Resolved: 6 years ago6 years ago
Resolution: --- → FIXED
[:jberkus] There is another bug logged for this that Brandon will fix today :: https://bugzilla.mozilla.org/show_bug.cgi?id=702359
> rsion=Firefox%3A8.0 talking about 133 crashes across multiple versions, when
> the link leading to it from the Firefox 8.0 topcrash report has 23 crashes
> and the "Reports" table also shows only those 23.

When I look at it, the Reports tab shows all 133.
(Reporter)

Comment 57

6 years ago
(In reply to [:jberkus] Josh Berkus from comment #56)
> > rsion=Firefox%3A8.0 talking about 133 crashes across multiple versions, when
> > the link leading to it from the Firefox 8.0 topcrash report has 23 crashes
> > and the "Reports" table also shows only those 23.
> 
> When I look at it, the Reports tab shows all 133.

On the first URL in comment #52, yes. On the second, no, it only shows the 23 ones for 8.0, as it should.
(Reporter)

Updated

6 years ago
Blocks: 702384
Kairo,

OK, that's bizarre.  What seems to happen is that if I navigate around a bit, the version= portion vanishes off the URL, and it stops filtering by version.
mbrandt discovered an issue with the Signature Summary, in that if a signature has quote marks in it, it doesn't display the data.

Example:

https://crash-stats.allizom.org/report/list?range_value=7&range_unit=days&date=2011-11-08&signature=arena_dalloc_small | arena_dalloc | free | nsPrefBranch%3A%3A`vector deleting destructor''(unsigned int)&version=Firefox%3A8.0

This is probably an issue with string escaping when building the query.
Ignore the preceding, the problem is something else.  Continuing discussion on bug #701815
Blocks: 701815
Ignore the preceding, the problem is something else.  Continuing discussion on bug #701815
Created attachment 576192 [details]
Signature summary tab

QA verified on stage.

Updated

6 years ago
Status: RESOLVED → VERIFIED
(Reporter)

Updated

6 years ago
Blocks: 705883
(Reporter)

Updated

6 years ago
Blocks: 711223
(Reporter)

Updated

6 years ago
Blocks: 711224
(Reporter)

Updated

6 years ago
Blocks: 711225
Component: Socorro → General
Product: Webtools → Socorro

Updated

6 years ago
Duplicate of this bug: 524851

Updated

6 years ago
Blocks: 524851
You need to log in before you can comment on or make changes to this bug.