Hi everyone, Thank you all for your feedback on the Alertmanager Time Interval design doc so far. The design has been greatly simplified and (in my opinion) improved already due to the feedback received.
I've made some changes to the proposed implementation so that it works on receivers now, not the routes themselves. Brian pointed out that this approach requires users to define many of their routes twice, with the only difference being the active time interval and which receiver to use. For example, notice how in the below snippet the 'severity: warning' alerts have two routes but really all that needs to change is the receiver and time: > - match: > severity: warning > time_intervals: > include: > - business_hours > exclude: > - public_holidays > receiver: team-X-pager > continue: true > - match: > severity: warning > receiver: team-X-slack > time_intervals: > exclude: > - business_hours So the current proposal is to introduce a 'timed_receivers' section of a route that pairs receivers and time intervals. This way the above route is simplified into a single block: > - match: > severity: warning > receiver: team-X-slack > timed_receivers: > - receiver: team-X-pager > include_intervals: > - business_hours > exclude_intervals: > - public_holidays > Additionally, this approach maintains the desirable characteristics of previous solutions in that it has no impact on existing routing decisions. Adding the tag also maintains backwards compatibility for existing configurations. Please let me know what you think, either here or in the comments of the design doc found here. <https://docs.google.com/document/d/1pf-rPDQUGJUHazyr5vanTO6ft3loNZO9UoVpvhShFtA/edit?usp=sharing> Cheers, Ben On Mon, Jun 8, 2020 at 1:31 PM Benjamin Ridley <[email protected]> wrote: > Hi everyone, > > I'm sure many of you have come across the problem of controlling alerts > based on the time of day or outside business hours etc inside Alertmanager. > There is also a longstanding issue > <https://github.com/prometheus/alertmanager/issues/876> on the > Alertmanager GitHub about this which I encourage you to read if you want > some more context. > > This is a proposed design for defining time intervals in the Alertmanager > configuration file and how they would be used in the routing tree to > silence particular routes inside or outside the specified intervals, > allowing users to model time-based requirements to their liking. > > The document is open for suggestions and comments and any feedback is > welcomed, so please take a look and let us know what you think. You can > access the document here > <https://docs.google.com/document/d/1pf-rPDQUGJUHazyr5vanTO6ft3loNZO9UoVpvhShFtA/edit?usp=sharing> > . > > Cheers, > Ben > > -- > You received this message because you are subscribed to the Google Groups > "Prometheus Developers" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to [email protected]. > To view this discussion on the web visit > https://groups.google.com/d/msgid/prometheus-developers/aa340edb-eb89-4dd8-af60-886d0427d3c7o%40googlegroups.com > <https://groups.google.com/d/msgid/prometheus-developers/aa340edb-eb89-4dd8-af60-886d0427d3c7o%40googlegroups.com?utm_medium=email&utm_source=footer> > . > -- You received this message because you are subscribed to the Google Groups "Prometheus Developers" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/prometheus-developers/CABaapwJNbHsHpcQ9WxH4XiJjWm17yv3G2HbP473JM1fJScF_tA%40mail.gmail.com.
