Add the Mozilla central servo sync account as an operator.

[larsbergstrom]

r? @edunham

@globau tells me that https://github.com/moz-servo-sync is the account that sheriff backouts will appear as. In order to auto-approve and support treeclosed, I was thinking that it would make sense to make it an operator.

---

This change is 

[edunham]

@larsbergstrom I agree with the idea of giving a shared sheriff account sufficient perms to close the tree. Can you or @globau point me to docs on:

• Who gets access to that account Especially, who hands out access when they decide that a new sheriff needs it?
• When they get access, what docs will a new account collaborator read to learn what they should and shouldn't do with it? (I'd like to make sure our instructions there are kept up to date if tree closure ever changes)
• When someone leaves, what's the process for deprovisioning them from the shared account?

Even if these docs are behind a Mozilla LDAP wall, I would like to look over them just to make sure we aren't accidentally introducing a huge vulnerability by adding this account.

[globau]

Who gets access to that account

the only people who have access are those who also have root access to the systems running servo-vcs-sync. ie. selected dev-ops and developers. this excludes sheriffs and contributors.

best to ask fubar about what happens with regards to staffing changes and access.

[larsbergstrom]

From further conversations with @globau, the credentials are stored in a wonderfully secure manner with extremely limited access to the underlying API key. I'm comfortable that the Firefox protections are at least as good as our own protection of the @bors-servo account :-)

[aneeshusa]

Is this a human account or a bot account? If the latter, are we planning to add another interface for closing the tree than writing GitHub comments to bors-servo?

[globau]

it's a bot account.

lars and i discussed a few times an out of band signalling mechanism such as SQS (it would help reduce race conditions) but it wasn't really viable to fit that into the schedule. if need be we can investigate that later.

[larsbergstrom]

@edunham @metajack Can we get this landed now? We need this ASAP for testing!

[metajack]

@bors-servo r+

I'd appreciate a pointer to the code be left here, or a page created in the wiki so people know where to start looking if something weird happens.

[bors-servo]

📌 Commit 817c124 has been approved by metajack

[bors-servo]

⌛ Testing commit 817c124 with merge 90c52cb...

[larsbergstrom]

The Gecko side backout code is here:
https://bugzilla.mozilla.org/show_bug.cgi?id=1365181

[bors-servo]

💔 Test failed - status-travis

[larsbergstrom]

@bors-servo retry

• repo timeout thing

[bors-servo]

⌛ Testing commit 817c124 with merge e013e00...

[bors-servo]

☀️ Test successful - status-travis
Approved by: metajack
Pushing e013e00 to master...

[aneeshusa]

I deployed this yesterday.