ratchet

README

ratchet

AI bot to help reduce operational toil

Code organization

• We are trying to keep things simple by only targetting one database, integration, etc.
• For backend we use postgres managed by sqlc. All that code is in internal/storage.
• Think of slack integration (at internal/slack) as a client and bot (at internal) as server.
  • Based on interactions that happens on slack, bot returns 0 or more actions to take.
  • This way we can write tests for the bot without messing with slack.
  • Later, we should consider replaying Slack HTTP API requests/responses to test this integration.

Database Schema

• The whole thing revolves around "Service" (to account for re-orgs)
• Service have alerts.
• Alerts have runbooks. All the past versions are kept in system, exactly one is active.
• Service has human asking for help. The whole slack thread is one instance of human interaction.
• Service has bot sending notifications about events related to it. Each notification is a separate instance.
• Alerts, humans and bot notifications come via one or more channels.
• Each channel is owned by a team. Owning team can change. Team can have multiple channels.
• Team names can be changed, teams can merge.
• Each human is either a member of a team or a customer asking for help.

Contributing

• To an slack workspace where you are admin, add an app using the manifest from app-manifest.yaml.
• Get access to a working Slack app/bot token and add it to .env (.gitignore'd) file as -

  RATCHET_SLACK_APP_TOKEN=xapp-...
  RATCHET_SLACK_BOT_TOKEN=xoxb-...

• Just start the binary using go run ./cmd/ratchet/main.go. It depends on docker to start a postgres instance.