1452940 - Add mozilla-esr60 and comm-esr60 to intermittent failures view (and maybe orange factor)

Status: RESOLVED DUPLICATE of bug 1449513

(Tree Management :: Treeherder, enhancement)

Description

[Johan Lorenzo [:jlorenzo]]

Firefox ESR 60.0 goes to build on April 30th. Like in bug 1337091, we want  these repo to appear on Orange Factor. Please note comm-esr60 hasn't been created yet.

Thank you!

Comment 1

[Justin Wood (:Callek)]

Ed, are you still the point person for this work?

Comment 2

[Ed Morley [:emorley]]

No, Geoff is. 

Though TBH anything to do with the legacy OrangeFactor should really be wontfix at this point.

Comment 3

[Justin Wood (:Callek)]

Hey Geoff,

I know this bug was cloned from what we needed to do for esr52, so (A) do we need this manually added for esr60 to OrangeFactor, (B) is there something manual to do for whatever is not "Legacy Orangefactor"

Comment 4

[Phil Ringnalda (:philor)]

No, OF shouldn't be wontfix. That's like saying "we replaced the grocery store with a shiny new vending machine which delivers perfectly done baked potatoes, so we should lock the doors of the grocery store immediately, while continuing to include weekly ad circulars in the newspaper for everything the grocery store used to sell, directing people to get those things from the baked potato vending machine, despite it only having baked potatoes."

What "replaced" OF was https://treeherder.mozilla.org/intermittent-failures.html, which covers only two "trees", Trunk and (for impenetrable reasons) Autoland. So we do weekly comments in bug 1429595 saying that it is either our most or second-most frequent failure, providing a link where you can see zero of the couple hundred failures. If, as they might well do, someone wants to know how frequently something is failing on esr60, or wants to find logs for failures there, their only hope is to remember that OF existed.

Comment 5

[Ed Morley [:emorley]]

Adding the other repositories to the new intermittent failures view is single digit line change to a constant in the Treeherder repository, the same as it is for OrangeFactor. The difference being that:
(a) it's putting effort into the new tool, and not one that is being switched off in a few months (due to the SCL3 datacenter decom)
(b) to deploy, someone doesn't have to SSH in and run a bunch of annoying manual steps

Comment 6

[Geoff Brown [:gbrown]]

So let's get this added to intermittent failures view (and other trees/repos while we are it) now and make that great!

We are really trying to put brasstacks away for good, but if it is still lingering a few weeks from now, let's reconsider an OF update.

Comment 7

[Geoff Brown [:gbrown]]

BTW, I notice bug 1337091 actually talks about "treestatus", which I have no knowledge of. Do we need another bug for updating treestatus?

Comment 8

[Henrik Skupin [:whimboo][⌚️UTC+2]]

(In reply to Geoff Brown [:gbrown] from comment #6)
> So let's get this added to intermittent failures view (and other trees/repos
> while we are it) now and make that great!

This is what bug 1449513 is about, so this bug should depend on it?

Comment 9

[Phil Ringnalda (:philor)]

treestatus is already done. Comment 0 meant "like bug 1337093"

Comment 10

[Sarah Clements [:sclements]]

FYI, in the meantime you can access any repo's failures via the query string param 'tree'.

Comment 11

[Sarah Clements [:sclements]]

I'll add repo's that are on OF to Intermittent Failures View. Are there any others that should be added? Is there a need to see fx-team and mozilla-aura?

Comment 12

[Phil Ringnalda (:philor)]

The only "trees" I ever used on OF were Trunk and All, so I can't say for sure, but I can't imagine any reason why anyone would want fx-team, mozilla-aurora, mozilla-esr45 or comm-esr45.

(Well, okay, after using OF with a wide date range to find out just how long-dead those long-dead trees really were, I actually did use the presence of a few non-closed bugs in their list of top failures to find things which should have been resolved WFM a year ago, but there are easier and better ways to find that information.)

Comment 13

[Phil Ringnalda (:philor)]

Oh, I'm days late and in the wrong bug, but also: despite still being in Treeherder's repo list for some reason (indifference being the most likely), comm-aurora is just as dead as mozilla-aurora, and can go too.

Comment 14

[Ed Morley [:emorley]]

I've filed bug 1458184 to disable comm-aurora (and others) in the rest of Treeherder.

Comment 15

[Sarah Clements [:sclements]]

Can all interested parties please confirm this as the final list, and I'll revise my patch (attached to bug 1449513):
'all',
'trunk',
'autoland',
'comm-beta',
'comm-central',
'comm-esr52',
'comm-esr60',
'mozilla-beta',
'mozilla-central',
'mozilla-esr52',
'mozilla-esr60',
'mozilla-inbound'

Comment 16

[Geoff Brown [:gbrown]]

lgtm

Comment 17

[Phil Ringnalda (:philor)]

wfm unless that's a sorted order, the order in which they will display, in which case I'd use

'all',
'trunk',
'mozilla-central',
'mozilla-inbound',
'autoland',
'mozilla-beta',
'mozilla-esr52',
'mozilla-esr60',
'comm-central',
'comm-beta',
'comm-esr52',
'comm-esr60'

Be nice for the future if we could identify who the interested parties actually are, and pin them down about their use. The lack of a mozilla-release on OF makes me suspect I'm not alone in not using individual trees, and we might well be able to cover every use with 'all', 'trunk', 'firefox-releases', 'firefox-esrs', 'comm-central', 'comm-releases'.

Comment 18

[Ed Morley [:emorley]]

(In reply to Phil Ringnalda (:philor) from comment #17)
> The lack of a mozilla-release on OF makes me suspect I'm not alone in not using individual
> trees, and we might well be able to cover every use with 'all', 'trunk',
> 'firefox-releases', 'firefox-esrs', 'comm-central', 'comm-releases'.

I agree with this. Duping over to bug 1449513 (since that's where the PR is), but I think we should use this suggestion instead.