Thanks for the report. That error almost always means something in your deployment config that Prefect expects to be an integer is being read as the string "None" (or a value with smart quotes like ‘None’), and Python is trying to convert it to an int. Common culprits in prefect.yaml/prefect_local.yaml: - schedule.interval (must be a number of seconds, not "None") - retries (must be an int) - retry_delay_seconds (must be an int) - any other numeric fields that were set to the string "None" instead of a real null Quick checks - Search your prefect_local.yaml for None or ‘None’ and change them to YAML nulls (null or ~) or remove the field entirely if you don’t need it. - Make sure you aren’t using smart quotes. Replace ‘ ’ with plain ' or remove quotes if not needed. - Ensure numeric fields are bare numbers, not quoted strings. Example If you have:

schedule:
  interval: 'None'
retries: 'None'
retry_delay_seconds: 'None'

Change to:

# remove schedule entirely if you don't want one
# or set a real interval in seconds
schedule:
  interval: 3600

# if you don’t want retries, remove the fields or set explicit ints
retries: 0
retry_delay_seconds: 0

A minimal working process-pool deployment example:

name: my-project
prefect-version: "3.*"

deployments:
- name: dep-name
  entrypoint: flows/my_flow.py:my_flow
  work_pool:
    name: process-pool
  # optional
  parameters: {}
  # optional schedule example
  # schedule:
  #   cron: "0 * * * *"
  #   timezone: "UTC"

A couple of questions to help pinpoint it: - What’s your Prefect version (

prefect version

)? - Can you share the relevant parts of your prefect_local.yaml (especially the deployments block and any schedule/retries fields)? - Is the exact error text showing fancy quotes around None (like ‘None’)? Docs for reference: prefect.yaml configuration in Prefect 3 Once you paste the snippet, I can call out the exact field causing the issue.

Thanks for the snippet — a couple things jump out that could cause this: 1) Invalid keys in Prefect 3.x -

tags

and

flow_name

are not valid deployment keys in Prefect 3.x

prefect.yaml

. The flow name is inferred from your

entrypoint

. Remove both of these keys. - Valid keys include

name

,

entrypoint

,

description

,

parameters

,

schedule

, and

work_pool

(plus

work_pool.job_variables

for per-run settings). - Docs: prefect.yaml in Prefect 3 2) The actual error: “‘None’” being parsed as an int - Your error shows curly quotes around None: ‘None’. That means somewhere in your YAML a value is literally the string ‘None’ (with smart quotes) for a field that expects an integer. - Common places this happens:

retries

,

retry_delay_seconds

,

concurrency

, or a

schedule.interval

. - Fix by removing those fields if you don’t use them, or set proper numbers, and ensure there are no smart quotes. Use

null

(or omit the field) instead of "None". Here’s a minimal, valid deployment block for a process pool you can drop into `prefect_local.yaml`:

name: my-project
prefect-version: "3.*"

deployments:
- name: my-service
  description: Run my service locally with Process
  entrypoint: app/flows/service_flow.py:my_service_flow
  work_pool:
    name: my-process-pool
    # optional job variables for process workers
    # job_variables:
    #   stream_output: true
    #   working_dir: .
    #   env:
    #     MY_ENV: "value"

If you intended to “tag” runs: deployment-level tags are not supported in 3.x. If you need labels on the worker job, you can use

job_variables.labels

(note: those are key/value labels for the worker, not run tags). Could you share: - Your full

prefect_local.yaml

(including the top-level `name`/`prefect-version` and any `schedule`/`retries` fields)? - Your Prefect version (

prefect version

)? With that I can point to the exact field that’s causing the

int()

on ‘None’.

A couple easy ways to get more detail (including stack traces) when running prefect deploy: One-off (just for this command)

PREFECT_LOGGING_ROOT_LEVEL=DEBUG prefect deploy --prefect-file prefect_local.yaml --name dep-name

Persist for your profile

prefect config set PREFECT_LOGGING_ROOT_LEVEL=DEBUG
# verify
prefect config view | grep PREFECT_LOGGING_ROOT_LEVEL

Optional: enable debug mode (if your Prefect version supports it)

prefect config set PREFECT_DEBUG_MODE=true
prefect config view | grep PREFECT_DEBUG_MODE

Then rerun your deploy command. If you still don’t see a full traceback, paste the DEBUG output here and I’ll help trace it to the exact field. Docs: - Logging customization - Settings reference