Create management interface

[adamcharnock]

What problems does this need to solve?

Use cases:

• Is everything working as expected?
  • Error rates
  • Services are behind
• If everything isn't working, why?
  • What are the recent errors?
  • Are services lacking workers?
  • Are workers bouncing? (keep restarting)
• Get me the message for a given ID (bonus: search for messages based on other criteria)
• Why was this message emitted? (i.e. does it have a parent event)
• What downstream events were triggered by this event?

---

Backend specific information:

• Redis:
  • Memory use
  • Stream size / growth over time
  • Last save to disk

---

Brain-dump of proposed features / possible interfaces:

• View services
• View workers
• View lag
• Inspection features as provided by lightbus inspect
• View APIs
• View events
• View RPC calls
• Send events (validated against schema)
• Make RPC calls (validated against schema)

Previous implementation

A previous implementation was created with aiohttp and React. I found this to be cumbersome, and likely a faff for other contributors to get up and running. I would like to prioritise ease of setup for the next attempt. I'm currently leaning towards Django Channels and vue.js.