What's the recommended best practice for Service naming conventions?


(Angela Apinyavat) #1

Some best practices for naming conventions would be the following:

  • Team names
  • Business Unit
  • Production environments
  • Product/feature
  • Priority level
  • Integration/Monitoring tool name
  • Customer name

For more information about reducing confusion with services and integrations by agreeing on naming conventions, see Integrations: Setting up services and integrations.

(michael gerber) #2

For all except the “Integration/Monitoring tool name”, we have a many-to-many relation between those items and monitoring systems. Some monitoring systems monitor OS level stats, some monitor networking, some monitor application services. Each team, business business unit, etc will need alerts from an overlapping set of monitoring tools. But as far as I can see, an integration can live in only one service, i.e. there is a one-to-many relation between services and integrations, not a many-to-many. That means “service” in PD terms is tied to integrations and we can’t use it for anything that does not align with integration i.e. monitoring system boundaries.
What entity can I use in PD to represent a service in the sense of an application or functional service?

(Angela Apinyavat) #3

Hi Michael,

A service can have multiple integrations. You can integrate with whichever monitoring tools you have managing your application to your PagerDuty Service. Hope that helps!

(michael gerber) #4

I stated that in my question (“there is a one-to-many relation between services and integrations”). What I am asking is whether one integration can belong to multiple services.

(Angela Apinyavat) #5

You would have to integrate the integration to multiple services. Each integration would have a unique integration key so that PagerDuty knows which service to route an incident to. Does that make sense?

(Rich Navis) #6

Updated Info… I’m not sure when Global Event Rules were introduced, but they allow you to use a single integration key, and route to services based on the payload of the event…