This work builds on the improvements we are making to Settings > Operations where we are moving all configuration related to Incident Management to Settings > Operations > Alerts OR Settings > Operations > Incidents.
This is one of the last settings we need to move.
Proposal
This issue will build on the work we're doing in #119018 (closed) and #119019 (closed), where we introduce a tab structure within the Incidents section on Settings > Operations. In this issue, we will move the Grafana integration into its own tab within the Incident section. That will look like this:
User can easily find all integration configuration in the same place making it more intuitive to learn about different options available for incident management.
@ameliabauerly I believe that the GitLab SRE team may start using the "auto-create issue from alerts functionality". I am wondering if rather than deprecating it, we move it somewhere else that makes more sense. People can still use it - if it is enabled then all alerts that come into GitLab and appear in the Alert list view will already have "incident" issues created for them and those will appear linked in the list view. If someone clicks on an alert ( to go to the detail page) then they will NOT see an option to create an issue. They will instead see a link that takes them to the Incident Issue.
Thoughts? This idea is new, not set in stone, so please give feedback. Any ideas on where the setting should live?
Thanks for the update, @sarahwaldner. If they are still going to use it, we definitely shouldn't deprecate it. But, if we are moving the alerts endpoint page to Settings > Operations, we might want to consider having two different categories on Settings > Operations: Incidents and Alerts. The alerts section could include the alerts endpoint configuration and any user-defined behaviors about what happens when an alert is received (ie, if we send an email when an alert is received, or if we create an issue). The incidents section could include, for example, configuring the Grafana integration so that Grafana graphs can be embedded in Incident issues. Just more clearly demarcating whether users are configuring alerts or incidents in each section. That's the work I was thinking this issue could tackle. Not sure how you feel about that?
@ameliabauerly Yes! I love that idea. I think that makes a lot of sense. That also gives us the opportunity in the future to add more rules/settings for how incidents are created from alerts in a central place. Please pursue this idea!!
Amelia Bauerlychanged title from Consider re-working incidents section on Settings > Operations to Move Grafana integration into Settings > Operations > Incidents
changed title from Consider re-working incidents section on Settings > Operations to Move Grafana integration into Settings > Operations > Incidents
Amelia Bauerlychanged the descriptionCompare with previous version
@ameliabauerly I want to move this to %13.5 - %13.4 is looking to be pretty full. Can you let me know if you have concerns about pushing this out one more milestone?
@ameliabauerly Is this critical to the user experience of this area? We have a lot of other high priority items that we need to tackle. I have moved it to the backlog but can revert that if you sfeel strongly that it needs to be addressed sooner than later.
@sarahwaldner - It's not critical. However, it would help smooth the process of setting up incidents and ensuring all the pieces needed to get incidents functioning properly are easy to access and all in one place.
Right now, everything else needed for incidents is within the incidents section. This is the only piece that's outside of it. It seems likely it will get missed if its separate. But, I understand that we probably have more important things to tackle currently. Hopefully we'll have a chance to get to this at some point soon, though, so all our incident settings stuff is in one place
@ameliabauerly That makes a lot of sense. Let's get a LOE from engineering.
@ClemMakesApps@crystalpoole Can you help me estimate LOE for this issue? While it is not a high priority, it is a major UX improvement that has been lagging for many milestones. If it is on the smaller side, we can complete it in %13.5 - otherwise we will push it out
@kbychu - Grafana embeds are used in incident investigation; they are posted in incident issues. That's why it's a little odd that they are currently listed separately from the other incident settings. The goal here is just to get all the setting items that impact incidents in a single place.
Do you not expect people to embed their Grafana charts beyond incident investigation? I can imagine other circumstances where this may be helpful, for example, a team proactively resolves disk running out of space and it would be useful to embed the chart in an issue. It seems to me the usecase is broader than just incidents, hence putting it under the incident setting doesn't make sense.
If we know people are not doing that at all, it would make sense to move it under incidents. Otherwise, I would leave as is until we have specific feedback.
Okay I think we should probably be optimizing the experience for incidents rather than for other, more hypothetical use cases. My focus is really ensuring that everything for incidents is as easy as possible to set up and get running. But, if your feeling is that we should wait to hear that this is a specific pain point for our users, then we can wait.
FYI, @sarahwaldner, sounds like we might not want to tackle this in 13.11. Not sure if you want to put it in the backlog?
@ohoral - I noticed you added yourself as an assignee - this may get pushed out depending on how the discussion above resolves. Just wanted to give you a heads up so you don't get started on something that we then stick in the backlog