Tasko, a self-hosted monitoring dashboard for taskiq (early WIP, looking for feedback)

[caushcani]

Hey folks 👋

I've been building Tasko, a broker-agnostic observability dashboard for
taskiq, task history, worker fleet health, queue metrics, without
wiring up Grafana/Prometheus yourself. Think something like Flower, but
built for taskiq's async-first, multi-broker world.

Repo: https://github.com/caushcani/tasko

What it does so far:

• Task explorer: filterable/sortable list, full args/kwargs/result/traceback
  on any run
• Worker fleet view: live heartbeats via a TaskiqMiddleware
• Overview dashboard (some panels still on placeholder data)

How it's built, in case it's interesting to anyone here:

• Broker support is a plugin architecture, tasko-redis, tasko-rabbitmq,
  etc. are separate packages implementing a small BrokerAdapter interface,
  discovered via entry points (same pattern taskiq itself uses for
  brokers). Redis is the only one implemented so far; RabbitMQ/NATS are
  stubs for later or for contributors, there's a
  guide for writing one
  if anyone's interested.
• Self-hosted only, by design, docker compose up, no SaaS.
• Next.js (SSR) + FastAPI + uv workspace, MIT licensed.

Where it's at: prototype stage, v1 scope is tasko-core +
tasko-middleware + tasko-redis + web. Dev setup and docker-compose
both work, uv sync && uv run tasko-core or the full docker stack, either
gets you a seeded dashboard in a couple minutes.

Genuinely curious:

• Does something like this already exist that I've missed?
• What's the "one thing" you'd want out of a taskiq monitoring tool that
  generic Prometheus/Grafana setups don't give you well?

Feedback, issues, and PRs all welcome, happy to talk through any of the
architecture if useful.