650953 - Improve the signal to noise ratio of wait notices to dev-tree-management

Status: RESOLVED WONTFIX

(Release Engineering :: General, defect)

Description

[Jeff Muizelaar [:jrmuizel]]

Currently we get 3 emails a day that are basically the equivalent to "everything's ok". These don't really provide value to most people that subscribe to dev-tree-management. It would be better if we only got emails when there were problems.

Comment 1

[John O'Duinn [:joduinn] (please use "needinfo?" flag)]

Being able to see when things are working well helps eliminate some confusion.  

Until we have better published trending dashboards for all our infrastructure, these posts to dev.tree-management are the only way people can see if our infrastructure is working smoothly, and keeping up with pending jobs over the day. In that situation, its not enough to only report problems - having confirmation that "all is ok" is useful data when people are trying to figure out whats going on. I therefore believe we should continue to publish these 3 reports, daily, even when things are going well, and think we should WONTFIX this bug. 

We setup dev.tree-management for this, and other automated RelEng status posts, that people didnt want polluting dev.planning or other groups. We intentionally setup the "wait time" posts so they can be easily filtered. If "wait times" are not of interest to you, hopefully you can filter these out.


Did that answer your original question?

Comment 2

[Jeff Muizelaar [:jrmuizel]]

How about at least combining the emails into 1 instead of 3?

Comment 3

[John O'Duinn [:joduinn] (please use "needinfo?" flag)]

(In reply to comment #2)
> How about at least combining the emails into 1 instead of 3?

1) We used to have this data coming from 3 separate masters, hence the 3 separate emails. I'll have to investigate to see if we can now generate this in one place.

2) A requirement from DamonS, Shaver, beltzner was to have this info in the subject field, so they could hit delete if the numbers looked good, and only open-to-read if the numbers looked bad.

tbh, I still think the right solution is to leave this as is for now, and keep working towards dashboards that have proper trending info of our load - so people could answer "are things working ok" by look at a graph/dashboard instead of wading through emails to a newsgroup.

Comment 4

[John O'Duinn [:joduinn] (please use "needinfo?" flag)]

(In reply to John O'Duinn [:joduinn] from comment #3)
> (In reply to comment #2)
> > How about at least combining the emails into 1 instead of 3?
> 
> 1) We used to have this data coming from 3 separate masters, hence the 3
> separate emails. I'll have to investigate to see if we can now generate this
> in one place.
> 
> 2) A requirement from DamonS, Shaver, beltzner was to have this info in the
> subject field, so they could hit delete if the numbers looked good, and only
> open-to-read if the numbers looked bad.
> 
> tbh, I still think the right solution is to leave this as is for now, and
> keep working towards dashboards that have proper trending info of our load -
> so people could answer "are things working ok" by look at a graph/dashboard
> instead of wading through emails to a newsgroup.

:jrmuizel: any objections to this proposal?

Comment 5

[Jeff Muizelaar [:jrmuizel]]

(In reply to John O'Duinn [:joduinn] from comment #4)
> This sounds reasonable.

Comment 6

[John O'Duinn [:joduinn] (please use "needinfo?" flag)]

(In reply to Jeff Muizelaar [:jrmuizel] from comment #5)
> (In reply to John O'Duinn [:joduinn] from comment #4)
> > This sounds reasonable.

Closing after irc with jrmuizel.