Bring your towel and join one of the fastest growing data communities. Welcome to our second-generation open source orchestration platform, a completely rethought approach to dataflow automation.

Prefect Community

I’m having an issue with prefect cloud. My agent is running and nothing looks wrong in the logs, but for some reason when I try to run a flow with the same label as the agent the flow doesn’t run because it says there are no agents with the label I’ve specified. Any debugging suggestions would be much appreciated!

Hey <@U02H33GCQCB>, could you show me the labels for the Flow and the agent? The most common issue I see is doing something like
```flow = Flow(
    "example",
    run_config=LocalRun(labels="dev")
)```
when it should be a list of strings
```flow = Flow(
    "example",
    run_config=LocalRun(labels=["dev"])
)```


Screen Shot 2021-12-06 at 9.22.21 AM.png

Agent:
```[2021-12-06 15:06:24,978] INFO - prism-acuity-aks-agent | Starting KubernetesAgent with labels ['prism-acuity']```
Flow:
```    cron_clock = clocks.CronClock(
        cron="0 */12 * * *", start_date=start_date, parameter_defaults=None, labels=["prism-acuity"],
    )```
I’m running the flow manually, and I see `prism-acuity` in the flow run UI

When you go to your `Agents` screen in the UI, do you see the agent you have running?

image.png

Yes, we had 2 agents with the same label for some reason
I removed the one that was 3 days stale

But the one whose logs I’m tailing still doesn’t seem to pick up the flow run

Do you have more than 700 late runs by chance?

I’m still a bit new to prefect. Not sure how I would even figure that out tbh

LoL it’s one of the ones I ran manually

I canceled that ^ and kicked off another manual run and it worked this time

This is pretty confusing. Seems like you ran Flows before too, so did this just stop working?

Ah ok that’s good. I think you should be good, but do ping if this still persists

Any idea what that phantom agent in the UI could have been?
What happens when a flow matches two agents it runs on both or it picks one?

Screen Shot 2021-12-06 at 10.42.34 AM.png

The UI just keeps stale agents. You can hit Clean UP in the top right. This is because if you turn on an agent and turn it on again with the same name, it takes the same place as the stale agent. So we keep track of them match future agents and continue that Run History

Those agents are likely off if the Last Queried is more than a few minutes

I think it’s something wrong with the flow
I’ve noticed that on the flow’s page in the UI I see the same “Label Problem” error
If I delete the label when I create a run and add it back again, then everything is good
Maybe there’s some whitespace in the label that isn’t rendering in the UI somehow?

I think whitespace should show. Could you try re-registering that? It is pretty weird

I re-registered it once this morning, but I can try again in a bit