这是indexloc提供的服务,不要输入任何密码
Skip to content
This repository was archived by the owner on Feb 13, 2025. It is now read-only.

Conversation

@pradeepbbl
Copy link
Contributor

@pradeepbbl pradeepbbl commented Jul 4, 2018

We have noticed bosun sending actionBodyClose instead of actionBodyDelayedClose when alerts are getting delayed closed. This is causing confusion because users are getting closed notifications immediately whereas it's still active in bosun, also when bosun actually closes the alert (after delay timeout) it's not sending any notifications and can only be viewed in the incident page.

In this pull request we try to solve both problems (sending delay close and notify when actual close happen), could you please review and let us know your thoughts?

…lose and trigger action notification

when delay close alert are getting `close`, `force close` & `cancel close`

      - cmd/bosun/web/web.go: added condition to check if the alert was delayed close
      - cmd/bosun/sched/check.go: trigger action notification when delayed close alert are getting `close`, `force close`, and `cancel close`
@pradeepbbl
Copy link
Contributor Author

@kylebrandt / @captncraig could you please review this pull request

@stepashka
Copy link

Hey @kylebrandt, @captncraig!
Any objections to this change? Is there something else we need to implement before this gets merged?
Thanks,
Anna

@kylebrandt
Copy link
Member

Probably need @captncraig to review the behavior on this one.

In terms of how the default behavior was supposed to go in my head, at least in regard to emails:

  • DelayedClose means "I'm pretty sure I have fixed the incident because I did x, y, z." The notification happens then because that is what other people (at least in my story I constructed :-/) need to know - the things has probably been dealt with and the monitoring system just needs to catch up.
  • From there, you can just assume it is closed, if the problem still exists after the close duration then a new incident will be created

At least one problem with the story is that notifications (at least how I have constructed them) don't give you any information about the previous incident on the same alertkey. So it the person getting the new incident might not know about the previous one.

I'm okay customizing this behavior I'm just worried about changing default behavior here and maybe adding some complexity. Basically the falls under the idea that "less notifications are better" - but if @captncraig thinks the complexity is fine I'm okay with being able to customize this behavior.

@pradeepbbl
Copy link
Contributor Author

@captncraig could you please share your thoughts on this?

@svagner
Copy link
Contributor

svagner commented Jan 15, 2020

@tdinucci, @muffix We are using this change more than 1 year already. Can you please review this MR?

@tdinucci
Copy link
Member

@svagner we will consider but as no test coverage cannot accept as it stands

@svagner
Copy link
Contributor

svagner commented Jan 16, 2020

@tdinucci , I agree there is really required some amount of tests. I'll make tests then:) Thanks!

@stale
Copy link

stale bot commented Jan 10, 2021

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

@stale stale bot added the wontfix label Jan 10, 2021
@stale stale bot closed this Feb 10, 2021
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.

Labels

Projects

None yet

Development

Successfully merging this pull request may close these issues.

5 participants