taskiq

Requires TaskIQ 0.11+ and <1, Python 3.10+. See the compatibility matrix for the full pin string.

Install

pip install z4j-taskiq

Wiring

taskiq’s extensibility is middleware-based. The framework adapters wire it automatically; for bare usage, use attach_to_broker or add the middleware manually:

from taskiq_redis import RedisAsyncResultBackend, ListQueueBroker
from z4j_taskiq import attach_to_broker

broker = ListQueueBroker("redis://localhost:6379")
broker = broker.with_result_backend(RedisAsyncResultBackend("redis://..."))
attach_to_broker(broker)

attach_to_broker adds Z4JTaskiqMiddleware and is the supported entry point.

What it captures

taskiq hook	z4j event
`pre_send`	`task_sent`
`pre_execute`	`task_started`
`post_execute` (no exception)	`task_succeeded`
`post_execute` (exception)	`task_failed`
retries (via retry middleware)	`task_retry`

Actions

Verb	How
`retry`	`task.kiq(args, *kwargs)` with original payload
`cancel`	`await broker.cancel_task(task_id)` where backend supports it
`purge_queue`	backend-specific; Redis: `FLUSHDB`-scoped

Multi-broker

taskiq supports multiple brokers in one process. Each broker appears separately in the dashboard.

Schedules

taskiq-scheduler is writable: schedules stored in Redis / Postgres / file backends can be created and modified from the UI. See scheduler: taskiq-scheduler.

Caveats

ZeroMQBroker - in-memory distribution between processes. Works with z4j but each process is a separate agent.
AioPikaBroker (RabbitMQ) - queue length via RabbitMQ management API.
InMemoryBroker - local-only, useful for tests; z4j captures events but there’s no cross-process visibility.

Multi-broker config

Frameworks that load multiple brokers (e.g. one Redis + one RabbitMQ) call attach_to_broker on each. Each broker appears as its own engine entry in the dashboard.