Short answer: Use

client.create_concurrency_limit

for tag-based limits. There is no

client.create_concurrency_limit_v2

method on the Python client—“v2” refers to the REST API for global concurrency limits, not tag-based limits. - Tag-based limits (tasks with specific tags):

client.create_concurrency_limit(tag=..., concurrency_limit=...)

- Global limits (general purpose/rate limiting across any code):

client.create_global_concurrency_limit(...)

Example (Prefect 3.x):

from prefect import get_client

async def setup_tag_limit():
    async with get_client() as client:
        await client.create_concurrency_limit(tag="database", concurrency_limit=5)

Notes: - Tasks must include the tag, e.g.,

@task(tags=["database"])

. - Under the hood, Prefect backs tag limits with a global limit named

tag:{your_tag}

. Docs: - How-to: Tag-based concurrency limits - Concepts: Tag-based concurrency limits - Client API reference - REST: Create tag-based concurrency limit If you were aiming for rate limiting or cross-task/flow resource control, use

client.create_global_concurrency_limit

instead. If you’re on Prefect 2.x, let me know and I can tailor the guidance.