Lester Cayabyab
07/16/2025, 6:30 PMflow.from_source(
source=GcsBucket.load(block_name), entrypoint="main.py:call_build"
).deploy(
name=build_name,
work_pool_name="default",
# cron=cron,
parameters={
"client_name": client_name,
"heartbeat_monitor_url": heartbeat_monitor_url,
},
)
I'm using "prefect[gcp]>=3.4.7"Marvin
07/16/2025, 6:30 PMMarvin
07/16/2025, 6:31 PMbash
prefect deployment ls
2. Then inspect your specific deployment:
bash
prefect deployment inspect "your-flow-name/build-name"
3. If you're not seeing the parameters, try explicitly setting the parameter schema in your flow definition:
python
from prefect import flow
@flow
def call_build(client_name: str, heartbeat_monitor_url: str):
# your flow code here
pass
4. You might also want to verify that the GCS bucket block is correctly configured and accessible:
python
from prefect_gcp.cloud_storage import GcsBucket
# Test block access
bucket = GcsBucket.load(block_name)
print(bucket.bucket_name) # Should print your bucket name
If after trying these steps you're still not seeing the parameters, could you please share:
1. The output of prefect deployment inspect
for your deployment?
2. The flow function definition (call_build
) from your main.py?
3. Any error messages you might be seeing?
This would help me provide more specific guidance for your situation.Lester Cayabyab
07/16/2025, 6:32 PM