Skip to main content

lightweight async distributed task queue

Project description

kuu

PyPI Python License Downloads

kuu


a native distributed task queue for python

uv add kuu
# extras: dashboard, nats, postgres, prometheus, redis

quick start

# myapp/app.py
from kuu import Kuu
from kuu.brokers.redis import RedisBroker
from kuu.results.redis import RedisResults

app = Kuu(broker=RedisBroker(url=...), results=RedisResults(url=...))


# myapp/tasks.py
from typing import TypedDict
from datetime import timedelta
from .app import app


class ChargeResult(TypedDict):
    ok: bool
    charged: int


@app.task
async def charge(user_id: int, amount_cents: int) -> ChargeResult:
    return {"ok": True, "charged": amount_cents}


@app.sched(every(hours=4, starting=time(hours=1, minutes=30))) # 1:30, 5:30, 9:30...
async def refresh_balance() -> None: ...


# myapp/main.py
from .tasks import charge


async def run() -> None:
    # type checker infers TaskHandle[ChargeResult]
    # args/kwargs of the task remain typed
    handle = await charge.q(user_id=1, amount_cents=500)

    # type checker infers ChargeResult
    result = await handle.result(timeout=30)
# reads ./kuunfig.toml or [tool.kuu] in ./pyproject.toml
# starts control plane with all presets spawned
uv run kuu start

# singular preset with dashboard / remote uplink
uv run kuu start --preset ...

what's inside

  • brokers: Redis Streams, NATS JetStream, in-memory (for tests)
  • scheduler: interval jobs (@app.every) and composable cron-like schedules (@app.sched)
  • middleware: logging, retry with exponential backoff + jitter, timeout, plus custom hooks
  • events: pub/sub signals for task lifecycle (task_enqueued .. task_dead)
  • serialization: JSON (msgspec), Msgpack, Pickle, with extensible type coercion via marshal
  • persistence: SQLite (zero-config) and PostgreSQL backends for run/log history
  • dashboard: Starlette+HTMX web UI with live worker/queue stats and task management
  • prometheus: multiprocess metrics with worker-side emitter and client-side middleware
  • hot reload: watch filesystem changes, restart worker pool on settled batches

config

put the block below into kuunfig.toml, or under [tool.kuu] in your pyproject.toml

[default] holds base values; each [presets.<name>] overrides only the fields you set. unset fields fall back to [default]. a flat config (no [default] wrapper) still works

[default]
queues = []              # consume from; empty = auto-discover from registry
processes = 1            # worker subprocesses to spawn
concurrency = 64         # max concurrent tasks per worker
prefetch = 16            # batch size; defaults to max(1, concurrency // 4)
shutdown_timeout = 30.0  # seconds to wait for in-flight tasks on stop

[default.metrics]
enable = false
host = "0.0.0.0"
port = 9191

[default.dashboard]
enable = false
host = "0.0.0.0"
port = 8181
path = "/dashboard"

scheduler.enable = false    # run scheduler loop in-process; jobs declared via app.every / app.sched

[default.watch]
enable = false              # reload workers on filesystem changes
root = "."                  # path to watch
respect_gitignore = true    # skip files matched by .gitignore
exclude = [".git/**"]       # extra globs to exclude
reload_delay = 0.25
reload_debounce = 0.5

[default.persistence]
enable = true               # store run/log history
dsn = "sqlite:///./kuu.db"  # sqlite (default) or postgres://...;
# also can be provided via KUU_PERSISTENCE_DSN env var
schema = ""                 # postgres schema; empty = default
runs_table = "kuu_runs"
logs_table = "kuu_run_logs"
keep_days = 7               # auto-purge runs older than this
max_runs = 100_000          # hard cap on stored runs
log_level = "INFO"
capture_args = true         # capture task args/kwargs in run detail

[presets.prod]
processes = 8
concurrency = 256

[presets.dev]
processes = 1
concurrency = 16

any setting can be overridden from the CLI with -o dotted.path=value values are parsed as JSON when possible (true, 42, ["a","b"]), otherwise kept as strings

contribution

well if you insist... (issues / PRs welcome)

#clankersgoaway

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

kuu-0.2.4.post1.tar.gz (89.3 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

kuu-0.2.4.post1-py3-none-any.whl (127.6 kB view details)

Uploaded Python 3

File details

Details for the file kuu-0.2.4.post1.tar.gz.

File metadata

  • Download URL: kuu-0.2.4.post1.tar.gz
  • Upload date:
  • Size: 89.3 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: uv/0.11.14 {"installer":{"name":"uv","version":"0.11.14","subcommand":["publish"]},"python":null,"implementation":{"name":null,"version":null},"distro":{"name":"Ubuntu","version":"24.04","id":"noble","libc":null},"system":{"name":null,"release":null},"cpu":null,"openssl_version":null,"setuptools_version":null,"rustc_version":null,"ci":true}

File hashes

Hashes for kuu-0.2.4.post1.tar.gz
Algorithm Hash digest
SHA256 84ffcf97ca2198b100f06d63cf06f708c54b6b6626d26bb5a53606a4e3318692
MD5 9262c23dc936170df9765f200133c3fd
BLAKE2b-256 42bb1b7bda27526ad5ac18c64639255b884c996e68f83904491c54ca48fb2cd8

See more details on using hashes here.

File details

Details for the file kuu-0.2.4.post1-py3-none-any.whl.

File metadata

  • Download URL: kuu-0.2.4.post1-py3-none-any.whl
  • Upload date:
  • Size: 127.6 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: uv/0.11.14 {"installer":{"name":"uv","version":"0.11.14","subcommand":["publish"]},"python":null,"implementation":{"name":null,"version":null},"distro":{"name":"Ubuntu","version":"24.04","id":"noble","libc":null},"system":{"name":null,"release":null},"cpu":null,"openssl_version":null,"setuptools_version":null,"rustc_version":null,"ci":true}

File hashes

Hashes for kuu-0.2.4.post1-py3-none-any.whl
Algorithm Hash digest
SHA256 66583fd058cf0fb1dea41164a6f1016756c9a2deaa4f33ccc5e95a6e4d006926
MD5 4a3d532dfe5608a61695560383998018
BLAKE2b-256 162991f8ab3335878070ce0f8574801575a0b982cbe5ed4bd4be33099e5a1730

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page