[mozproy] Drop zero-sized patches from recording at replay time
Categories
(Testing :: Mozbase, defect, P2)
Tracking
(Not tracked)
People
(Reporter: alexandrui, Assigned: alexandrui)
References
Details
(Keywords: perf-alert)
Attachments
(1 obsolete file)
Currently the recordings are capturing the zero-sized patches that sometimes cause intermittent failures by sometimes waiting on patches that has no content. We need to drop those to avoid causing intermittents in the future.
Assignee | ||
Comment 1•4 years ago
|
||
Updated•4 years ago
|
Pushed by gmierz2@outlook.com: https://hg.mozilla.org/integration/autoland/rev/ece60a1dbd9d Drop zero-sized patches from recording at replay time. r=sparky,perftest-reviewers
Comment 3•4 years ago
|
||
bugherder |
Assignee | ||
Comment 4•4 years ago
|
||
Sparky, aren't the mb/mbu
fails related to the patch? I did another try to make sure of that.
Assignee | ||
Updated•4 years ago
|
Comment 6•4 years ago
|
||
Backed out for performance regressions (bug 1676630)
Backout: https://hg.mozilla.org/integration/autoland/rev/88f6a6c2a328a0abc4421e5428b9b1c7a527c8b3
Assignee | ||
Comment 7•4 years ago
|
||
I will investigate and come back with a fixed patch.
Comment 8•4 years ago
|
||
(In reply to Alexandru Ionescu (needinfo me) [:alexandrui] from comment #7)
I will investigate and come back with a fixed patch.
I would like to understand the issue and the impact it's having before we invest further time into this bug. Could you provide some more details?
Assignee | ||
Comment 9•4 years ago
|
||
I am having hard times (again) running browsertime locally. As soon as I fix them, I'll run the tests locally and investigate what's happening.
Comment 10•4 years ago
|
||
(In reply to Alexandru Ionescu (needinfo me) [:alexandrui] from comment #9)
I am having hard times (again) running browsertime locally. As soon as I fix them, I'll run the tests locally and investigate what's happening.
This doesn't answer my question. Once we understand the original issue and the impact we can prioritise this appropriately. I don't want to spend time investigating/fixing this patch if it doesn't provide enough value relative to the effort.
Comment 11•4 years ago
|
||
Backout got merged: https://hg.mozilla.org/mozilla-central/rev/88f6a6c2a328
Assignee | ||
Comment 12•4 years ago
|
||
I'll try to understand why is it impacting visual metrics this way as I'm very curious, but with low priority.
Assignee | ||
Comment 13•4 years ago
•
|
||
(In reply to Andreea Pavel [:apavel] from comment #6)
Backed out for performance regressions (bug 1676630)
Backout: https://hg.mozilla.org/integration/autoland/rev/88f6a6c2a328a0abc4421e5428b9b1c7a527c8b3
== Change summary for alert #27627 (as of Fri, 13 Nov 2020 02:29:23 GMT) ==
Regressions:
Ratio | Suite | Test | Platform | Options | Absolute values (old vs new) |
---|---|---|---|---|---|
26% | raptor-tp6-twitter-firefox-cold | loadtime | windows7-32-shippable | 1,427.17 -> 1,794.33 | |
25% | raptor-tp6-twitter-firefox-cold | loadtime | windows10-64-shippable | 1,441.58 -> 1,797.42 | |
24% | raptor-tp6-twitter-firefox-cold | loadtime | windows10-64-shippable-qr | webrender | 1,441.83 -> 1,793.08 |
8% | raptor-tp6-reddit-firefox-cold | loadtime | windows10-64-shippable-qr | webrender | 2,752.50 -> 2,979.92 |
8% | raptor-tp6-reddit-firefox-cold | loadtime | windows10-64-shippable | 2,737.50 -> 2,961.75 | |
8% | raptor-tp6-reddit-firefox-cold | loadtime | windows7-32-shippable | 2,709.08 -> 2,927.67 | |
6% | raptor-tp6-twitter-firefox-cold | windows10-64-shippable | 454.24 -> 483.12 | ||
6% | raptor-tp6-twitter-firefox-cold | windows7-32-shippable | 457.21 -> 483.95 | ||
5% | raptor-tp6-twitter-firefox-cold | windows10-64-shippable-qr | webrender | 450.16 -> 474.49 | |
3% | raptor-tp6-reddit-firefox-cold | windows7-32-shippable | 644.61 -> 661.35 |
Improvements:
Ratio | Suite | Test | Platform | Options | Absolute values (old vs new) |
---|---|---|---|---|---|
115% | raptor-tp6-paypal-firefox-cold | replay-confidence | windows10-64-shippable | 45.17 -> 97.00 | |
113% | raptor-tp6-paypal-firefox-cold | replay-confidence | windows10-64-shippable-qr | webrender | 45.50 -> 97.00 |
111% | raptor-tp6-paypal-firefox-cold | replay-confidence | windows7-32-shippable | 45.92 -> 97.00 | |
34% | raptor-tp6-instagram-firefox-cold | replay-confidence | windows10-64-shippable-qr | webrender | 71.83 -> 96.00 |
33% | raptor-tp6-instagram-firefox-cold | replay-confidence | windows10-64-shippable | 72.00 -> 96.00 | |
33% | raptor-tp6-instagram-firefox-cold | replay-confidence | windows7-32-shippable | 72.17 -> 96.00 | |
30% | raptor-tp6-twitch-firefox-cold | replay-confidence | windows7-32-shippable | 72.00 -> 93.25 | |
27% | raptor-tp6-twitch-firefox-cold | replay-confidence | windows10-64-shippable-qr | webrender | 72.00 -> 91.58 |
27% | raptor-tp6-twitch-firefox-cold | replay-confidence | windows10-64-shippable | 73.83 -> 93.58 | |
25% | raptor-tp6-yahoo-mail-firefox-cold | replay-confidence | windows10-64-shippable | 62.25 -> 78.08 | |
25% | raptor-tp6-yahoo-mail-firefox-cold | replay-confidence | windows7-32-shippable | 62.08 -> 77.58 | |
25% | raptor-tp6-yahoo-mail-firefox-cold | replay-confidence | windows10-64-shippable-qr | webrender | 62.75 -> 78.33 |
12% | raptor-tp6-netflix-firefox-cold | replay-confidence | windows10-64-shippable | 76.58 -> 85.67 | |
11% | raptor-tp6-netflix-firefox-cold | replay-confidence | windows7-32-shippable | 79.00 -> 88.00 | |
11% | raptor-tp6-netflix-firefox-cold | replay-confidence | windows10-64-shippable-qr | webrender | 76.50 -> 85.08 |
11% | raptor-tp6-reddit-firefox-cold | replay-confidence | windows7-32-shippable | 85.67 -> 94.92 | |
10% | raptor-tp6-reddit-firefox-cold | replay-confidence | windows10-64-shippable-qr | webrender | 85.83 -> 94.58 |
10% | raptor-tp6-reddit-firefox-cold | replay-confidence | windows10-64-shippable | 85.92 -> 94.33 | |
9% | raptor-tp6-netflix-firefox-cold | fcp | windows10-64-shippable-qr | webrender | 330.50 -> 301.67 |
8% | raptor-tp6-youtube-firefox-cold | fcp | windows10-64-shippable | 515.31 -> 476.08 | |
7% | raptor-tp6-amazon-firefox-cold | replay-confidence | windows10-64-shippable-qr | webrender | 80.20 -> 85.58 |
7% | raptor-tp6-youtube-firefox-cold | fcp | windows10-64-shippable-qr | webrender | 493.58 -> 460.58 |
6% | raptor-tp6-youtube-firefox-cold | fcp | windows7-32-shippable | 515.35 -> 482.00 | |
6% | raptor-tp6-linkedin-firefox-cold | replay-confidence | windows10-64-shippable | 85.33 -> 90.67 | |
6% | raptor-tp6-yahoo-news-firefox-cold | replay-confidence | windows10-64-shippable-qr | webrender | 86.00 -> 91.25 |
6% | raptor-tp6-yahoo-news-firefox-cold | replay-confidence | windows7-32-shippable | 86.00 -> 91.25 | |
6% | raptor-tp6-yahoo-news-firefox-cold | replay-confidence | windows10-64-shippable | 86.00 -> 91.17 | |
6% | raptor-tp6-amazon-firefox-cold | replay-confidence | windows7-32-shippable | 80.90 -> 85.67 | |
6% | raptor-tp6-linkedin-firefox-cold | replay-confidence | windows7-32-shippable | 85.08 -> 90.00 | |
6% | raptor-tp6-linkedin-firefox-cold | replay-confidence | windows10-64-shippable-qr | webrender | 85.83 -> 90.58 |
5% | raptor-tp6-youtube-firefox-cold | windows10-64-shippable-qr | webrender | 968.49 -> 920.64 | |
5% | raptor-tp6-youtube-firefox-cold | windows7-32-shippable | 989.73 -> 942.20 | ||
4% | raptor-tp6-youtube-firefox-cold | replay-confidence | windows7-32-shippable | 92.42 -> 96.00 | |
4% | raptor-tp6-youtube-firefox-cold | replay-confidence | windows10-64-shippable-qr | webrender | 92.50 -> 96.00 |
4% | raptor-tp6-youtube-firefox-cold | replay-confidence | windows10-64-shippable | 92.69 -> 96.17 | |
4% | raptor-tp6-google-firefox-cold | replay-confidence | windows7-32-shippable | 83.00 -> 86.00 | |
4% | raptor-tp6-microsoft-firefox-cold | replay-confidence | windows7-32-shippable | 74.89 -> 77.58 | |
3% | raptor-tp6-google-firefox-cold | replay-confidence | windows10-64-shippable | 83.00 -> 85.75 | |
3% | raptor-tp6-google-firefox-cold | replay-confidence | windows10-64-shippable-qr | webrender | 82.80 -> 85.50 |
3% | raptor-tp6-imdb-firefox-cold | replay-confidence | windows10-64-shippable | 91.00 -> 93.33 | |
3% | raptor-tp6-yandex-firefox-cold | replay-confidence | windows7-32-shippable | 82.00 -> 84.08 | |
3% | raptor-tp6-pinterest-firefox-cold | replay-confidence | windows10-64-shippable | 80.00 -> 82.00 | |
3% | raptor-tp6-pinterest-firefox-cold | replay-confidence | windows10-64-shippable-qr | webrender | 80.00 -> 82.00 |
3% | raptor-tp6-pinterest-firefox-cold | replay-confidence | windows7-32-shippable | 80.00 -> 82.00 | |
2% | raptor-tp6-yandex-firefox-cold | replay-confidence | windows10-64-shippable | 82.00 -> 84.00 | |
2% | raptor-tp6-yandex-firefox-cold | replay-confidence | windows10-64-shippable-qr | webrender | 82.00 -> 84.00 |
2% | raptor-tp6-imdb-firefox-cold | replay-confidence | windows10-64-shippable-qr | webrender | 90.84 -> 93.00 |
2% | raptor-tp6-imdb-firefox-cold | replay-confidence | windows7-32-shippable | 90.80 -> 92.92 | |
2% | raptor-tp6-facebook-redesign-firefox-cold | replay-confidence | windows10-64-shippable-qr | webrender | 87.00 -> 88.83 |
2% | raptor-tp6-outlook-firefox-cold | replay-confidence | windows10-64-shippable | 98.00 -> 100.00 | |
2% | raptor-tp6-outlook-firefox-cold | replay-confidence | windows10-64-shippable-qr | webrender | 98.00 -> 100.00 |
2% | raptor-tp6-outlook-firefox-cold | replay-confidence | windows7-32-shippable | 98.00 -> 100.00 |
For up to date results, see: https://treeherder.mozilla.org/perfherder/alerts?id=27627
Assignee | ||
Updated•4 years ago
|
Assignee | ||
Updated•4 years ago
|
Comment 14•3 years ago
|
||
[Tracking Requested - why for this release]:
Comment 15•3 years ago
|
||
(In reply to Florin Strugariu [:Bebe] (needinfo me) from comment #14)
[Tracking Requested - why for this release]:
Please explain why you're requesting tracking. It's not clear to me the issue and impact here.
Comment 16•3 years ago
|
||
I believe this was a mistake but :bebe can correct me if I'm wrong. This bug caused a performance regression when it landed leading to it being backed out. When we file independant regression bugs we set the status flag for the affected version. We likely need to improve our workflow for improvements and backouts as setting the flags on this bug doesn't make a lot of sense.
Comment 17•3 years ago
|
||
You are right. In my view we should not change the status flag on bugs that are not linked directly to performance alerts.
As we decided to only add a comment in the causing bug for improvements and not log a new bug for it. and we are tracking these status flags we should find a improvement in our process for these cases
Assignee | ||
Comment 18•3 years ago
|
||
Closing this as incomplete as the recording that was failing because of zero-sized patches is not failing anymore thus there's nothing we can test on to make sure the patch is doing what's supposed to and not bringing regressions with it.
The solution would be to drop the exact zero-sized patches with the exact status code(s) that are causing the issue.
Code to start with is attached to this bug and to Bug 1628627.
Assignee | ||
Updated•3 years ago
|
Updated•3 years ago
|
Updated•3 years ago
|
Description
•