intercom-archive / banjaxed Goto Github PK

I'm a big fan of feeds and with tools like Slack supporting feed subscriptions in particular channels I think it'd be very useful to have a feed of updates for Banjaxed incidents.

We could start small with a feed for all updates and think big with feeds based on severity level or if we ever expand into splitting incidents by service affected then feeds to those.

https://devcenter.heroku.com/articles/heroku-button

This would make it easier for someone to quickly set up the app and play around with it.

It'd be nice to take a look at the UI without having to spin up an instance.

Github's drag & drop feature is pretty cool. We're going to be linking to all sorts of graphs and uploading them somewhere else. I'm interested in discussing options here

/cc https://github.com/intercom/intercom/issues/18534

After action reports - the ability to require an after action report for an incident and for it to email the request to the person who the AAR was assigned to and to notify them every X days until the AAR is marked as DONE.

Unable to log into a heroku deployment without a valid GITHUB_ORG
For your typical end user, setting up an ORG is a pain and shouldn't be required.
Let's make this optional?

Reminders - notify owners and relevant -update email lists of incidents are open and have not been updated/mitigated/closed within the last X mins.

Slack Integration - The ability to overlay slack logs from selected channels for particular time periods on top of banjaxed comments. By default the time periods should be something like “unless otherwise specified, include slack logs from the selected channels from 10 mins before the incident started to 30 mins after the last comment in the incident”.

It would be really cool if PagerDuty incidents could automatically open banjaxed incidents.

This could be implemented with PagerDuty service-level webhooks, and would allow only certain services (say, high-sev ones) to open these incidents.

MVP: PagerDuty webhook opens new Banjaxed incident

Nice-to-haves:

• If a user exists in both tools, an acknowledgement in PagerDuty assigns it to that user in banjaxed.
• If a PagerDuty incident is resolved, the banjaxed incident closes.
• If a comment is entered in banjaxed, it is posted to the PagerDuty incident as a note.

Notifications - the theme here is that we want to make it easy for team member to stay in the loop on emerging and ongoing incidents without needing to build a full on banjaxed mobile app (though that’s definitely a possibility for the future).
⁃ banjaxed-events-open-all@ = an email list that get’s emailed when any new incident is open
⁃ banjaxed-events-updates-all@ = an email list that get’s email whenever an incident is updated
⁃ banjaxed-PX-(open|updates)@ = versions of the above for different priority incidents

Explanation of the intended workflow and a differentiation between each state. Eg whats the difference between resolved and closed?

Maybe chance the severity to P1, P2, P3, P4 or Sev1, Sev2, Sev3 or Sev4 (as they are more industry standard and make it configurable for people add/customise definitions for what each severity level mean)?

After action report - the ability to create and store after action reports in the tool.

Example incident where this would be really appreciated: http://banjaxed.internal.intercom.io/incidents/8

Would hugely increase readability. We could use Github flavoured markdown and keep consistent MD syntax from everywhere else we post comments / code https://github.com/github/markup

cc/ @eugeneius

Comment editing in Banjaxed isn't currently possible. It'd definitely be a nice feature to have. Perhaps we could go even bigger on this feature with an edit history to make sure we're not removing pertinent history

It would be useful to see how an incident evolved over time, not just its current state. When an incident's status changes, or a comment is edited, we should capture that information and make the log accessible.

Have banjaxed incidents created by a monitoring and alerting platform (pager duty, new relic, syren, pingdom, etc). We’re not using banjaxed enough, and we’re missing the opportunity to collect info about outages, so we want to auto open issue for people and then we also want to encourage people to keep them updated.

my initial thought is to create a webhook for opening incidents