The second big improvement affects crash pings and comparing signatures between crash pings and crash reports. Lots of groups are doing things with crash ping data. I regularly see comments regarding analysis comparing crash ping signatures to crash report signatures leading to conclusions about visibility of certain kinds of crashes.
Is anyone maintaining a "things to know about crash ping data" doc somewhere? This big change should be documented in it. On Thu, Sep 26, 2024 at 7:41 AM Gabriele Svelto <[email protected]> wrote: > [cross-posting to dev-platform and crash-reporting-wg] > > A few important improvements in crash reporting have landed recently and > it's worth highlighting them since they'll have a significant impact. > > First of all we've updated the minidump-writer crate, picking up several > months worth of work. Generating minidumps on Linux and Android now > takes hundreds of microseconds to a few milliseconds at most, a 25-30x > speedup which will be particularly welcome on slow and loaded machines > (low-end phones in particular). Additionally the logic that retrieves > information about modules does not rely on the files being available on > disk anymore. This will greatly reduce the number of cases where we're > left with truncated stack traces or missing symbols. Last but not least > some common failures when generating minidumps on Android are now > non-critical, so we expect more crashes to be captured successfully. > > The second big improvement is that the new Rust-based minidump-analyzer > has just landed. This is a tool that runs on the user machine after a > crash to populate our telemetry with stack traces. The previous version > of the tool, based on Breakpad, produced decent traces on 64-bit Windows > but poor ones on all other platforms. The new tool produces quality > stack traces on all platforms we support (and some we don't). These will > soon be fed into our new telemetry dashboards, providing a clearer > picture of Firefox stability and a better way to identify the most > significant issues. > > I'm taking this occasion to thank Alex Franchuk for the large amount of > work poured into reaching these milestones and Jake Shadle for his very > appreciated contributions. > > Gabriele Svelto > > -- > You received this message because you are subscribed to the Google Groups > "crash-reporting-wg" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to [email protected]. > To view this discussion on the web visit > https://groups.google.com/a/mozilla.com/d/msgid/crash-reporting-wg/0f6df3c9-774d-4b36-a556-67958a5fbee3%40mozilla.com > . > For more options, visit https://groups.google.com/a/mozilla.com/d/optout. > -- You received this message because you are subscribed to the Google Groups "[email protected]" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web visit https://groups.google.com/a/mozilla.org/d/msgid/dev-platform/CAKnh9qiDSKXmw9HWU6Em%2B6-QByWYW-%3DTp_OYy-BC2BoKJEsOQQ%40mail.gmail.com.
