Add a tool to request info when a bug is moved to Core::Performance

[suhaibmujahid]

Resolves #1887

Checklist

• Type annotations added to new functions
• Docs added to functions touched in main classes
• Dry-run produced the expected results
• The to-be-announced tag added if this is worth announcing

[marco-c]

What if the information is already present in the bug?

[suhaibmujahid]

What if the information is already present in the bug?

In that scenario the needinfo will be cleared with/without a confirmation that the information is already in the bug.

We could add a sentence at the end to request a response: "If the requested information is already in the bug, please confirm that."

@julienw wdyt?

[julienw]

Yeah, this sounds good to me. We can adjust later if we see this case happening too much.

[marco-c]

Could we add some basic checks to prevent some obvious cases? E.g. if an attachment with "memory" in its name is in the bug, we don't ask for a memory report; if there is a "share.firefox.dev" in any of the comments, we don't ask for a performance profile.

[julienw]

This would work for me too :-)

[suhaibmujahid]

Could we add some basic checks to prevent some obvious cases? E.g. if an attachment with "memory" in its name is in the bug, we don't ask for a memory report; if there is a "share.firefox.dev" in any of the comments, we don't ask for a performance profile.

Nice idea! However, doing simple checks could have high false negatives. I suggest monitoring the cases after deploying this and adjusting later if this introduces much noise.

I filed an issue to follow up on this idea (i.e., #1926).

@marco-c wdyt?

[marco-c]

I would add some simple checks now, the ones that are very unlikely to have false negatives and are easy to implement.

• "share.firefox.dev" in the comments is definitely a link to a profile, so it is very unlikely to cause false negatives.
• if you go to about:memory and click on "Measure and save...", Firefox will save a memory-report.json.gz file on your disk, so "memory-report.json.gz" in the attachments is also very unlikely to cause false negatives.

[suhaibmujahid]

Dry-run

2023-03-10 23:31:06,393 - INFO - Run tool moved_to_performance.py
2023-03-10 23:31:07,932 - INFO - Tool moved_to_performance.py has finished.

The following bugs were moved into the Core::Performance component; a comment to request more information was added:

Bug	Summary
1821555	Both FB Marketplace & Twitter lock up for 30-45 seconds every time I try to scroll.
1821424	Memory leakage 110.0.01 Win11 22H2
1820078	I try to access https://mysafaricost.com on firefox browser and it redirects always on some pages and stops and displays a broken site at the end while it opens on another browser.
1795009	Open multiple tabs becomes slow

Comment example - bug 1795009

This bug was moved into the Performance component.

:zero.arst, could you make sure the following information is on this bug?

• ✅ For slowness or high CPU usage, capture a profile with http://profiler.firefox.com/, upload it and share the link here.
• For memory usage issues, capture a memory dump from about:memory and attach it to this bug.
• Troubleshooting information: Go to about:support, click "Copy raw data to clipboard", paste it into a file, save it, and attach the file here.

If the requested information is already in the bug, please confirm that.

Thank you.

[suhaibmujahid]

@julienw in the context of doing some checks to avoid asking for information that is already present in the bug, do you think we should consider only content that was submitted be the reporter?

[marco-c]

I never used a checklist on Bugzilla, does it actually work? Can a user mark something in a checklist as done? Do they need specific permissions to do that if the comment is from somebody else (the bot in this case)?

[julienw]

@julienw in the context of doing some checks to avoid asking for information that is already present in the bug, do you think we should consider only content that was submitted be the reporter?

Yes, this sounds reasonable.
I wonder if we should also have a time limit. For example if the information is older than say 6 months, we could assume this is outdated and still ask about it.

[suhaibmujahid]

I wonder if we should also have a time limit. For example if the information is older than say 6 months, we could assume this is outdated and still ask about it.

Done in 2a7cae1

[suhaibmujahid]

I never used a checklist on Bugzilla, does it actually work? Can a user mark something in a checklist as done? Do they need specific permissions to do that if the comment is from somebody else (the bot in this case)?

Unfonratintly, Bugzilla does not support markdown checklists.

[marco-c]

Maybe when the data is present but not recent enough, we should adjust the comment to ask to refresh the info (otherwise people will ignore it and say that it is already present)

[suhaibmujahid]

Maybe when the data is present but not recent enough, we should adjust the comment to ask to refresh the info (otherwise people will ignore it and say that it is already present)

Do you think 13820a7 would be enough? If not, I can implement what you suggested.