1498609 - Add the MemoryErrorCorrection annotation to the crash ping

Status: RESOLVED FIXED

(Toolkit :: Crash Reporting, enhancement)

Description

[Gabriele Svelto [:gsvelto]]

This would be useful both for statistics (e.g. crash rate of ECC vs non-ECC machines) and to identify valid low-volume crashes. One of the issues with low-volume crashes is that they might be caused by memory errors - especially permanent ones - having this field would allow us to more easily weed out crashes due to hardware faults from valid ones.

Comment 1

[(not currently active) Ted Mielczarek]

I was actually thinking this might be valuable to have in the general telemetry ping as well. That might give us insight into whether ECC memory is noticeably different in our crashing population vs. general population.

Comment 2

[Gabriele Svelto [:gsvelto]]

Attachment: Bug 1498609 - Annotate crash pings to include ECC memory support (or lack of)

Comment 3

[Gabriele Svelto [:gsvelto]]

(In reply to Ted Mielczarek [:ted] [:ted.mielczarek] from comment #1)
> I was actually thinking this might be valuable to have in the general
> telemetry ping as well. That might give us insight into whether ECC memory
> is noticeably different in our crashing population vs. general population.

I'll open a separate bug for that. I'll get the data collection request reviewed here so that I won't need it anymore to add it to the main ping.

Comment 4

[Gabriele Svelto [:gsvelto]]

Attachment: Data review request

Comment 5

[Chris H-C :chutten]

(In reply to Gabriele Svelto [:gsvelto] from comment #3)
> (In reply to Ted Mielczarek [:ted] [:ted.mielczarek] from comment #1)
> > I was actually thinking this might be valuable to have in the general
> > telemetry ping as well. That might give us insight into whether ECC memory
> > is noticeably different in our crashing population vs. general population.
> 
> I'll open a separate bug for that. I'll get the data collection request
> reviewed here so that I won't need it anymore to add it to the main ping.

Welll... Depends if you're materially changing the collection. Adding an annotation to (infrequent) "crash" pings is a separate concern from adding something to the core Telemetry Environment, or adding a bool scalar only present on the first subsession of a session, or accumulating at regular intervals samples into a boolean histogram, or...

Give me a ping when you're settled on an implementation strategy for putting this in the "main" ping. It might be covered, but it might not be :)

Comment 6

[Gabriele Svelto [:gsvelto]]

(In reply to Chris H-C :chutten from comment #5)1
> Welll... Depends if you're materially changing the collection. Adding an
> annotation to (infrequent) "crash" pings is a separate concern from adding
> something to the core Telemetry Environment, or adding a bool scalar only
> present on the first subsession of a session, or accumulating at regular
> intervals samples into a boolean histogram, or...

OK, then consider this only for the crash ping.

> Give me a ping when you're settled on an implementation strategy for putting
> this in the "main" ping. It might be covered, but it might not be :)

Sure, it's probably something that would belong in the telemetry environment where we put all the other hardware info.

Comment 7

[Chris H-C :chutten]

Comment on attachment 9017134 [details]
Data review request

DATA COLLECTION REVIEW RESPONSE:

    Is there or will there be documentation that describes the schema for the ultimate data set available publicly, complete and accurate? 

The annotation is documented in the "crash" ping documentation.

    Is there a control mechanism that allows the user to turn the data collection on and off? 

Yes. Standard Telemetry mechanisms apply.

    If the request is for permanent data collection, is there someone who will monitor the data over time?

Yes, :gsvelto

    Using the category system of data types on the Mozilla wiki, what collection type of data do the requested measurements fall under?

Category 1.

    Is the data collection request for default-on or default-off?

Default on.

    Does the instrumentation include the addition of any new identifiers (whether anonymous or otherwise; e.g., username, random IDs, etc. See the appendix for more details)?

Yes, sorta. The string identifier of the type of memory error correction algorithm is now reported, and there is no central clearing house I can find that limits this or ensures it is read-only. It is unlikely to be or become an identifier, but it appears to be capable of doing so.

    Is the data collection covered by the existing Firefox privacy notice? 

Yes.

    Does there need to be a check-in in the future to determine whether to renew the data? 

No.

---
Result: datareview+

Comment 8

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/e72b6abf3ec4
Annotate crash pings to include ECC memory support (or lack of) r=ted

Comment 9

[Stefan Hindli [:stefan_hindli]]

https://hg.mozilla.org/mozilla-central/rev/e72b6abf3ec4