Can't get the reason why there are so many late ru...
# prefect-community
k
Can't get the reason why there are so many late runs piling up Have scheduled about 6-7 flows on single node cluster via minikube
Sharing the deployment scripts
k
Did you set up a workqueue and agent?
k
Have 3 work queues running
k
How about agent?
k
Running on a single agent created via manifest file
k
Which work queue is that agent using?
k
Copy code
- name: agent
          image: prefecthq/prefect:2.0b5-python3.9
          command: ["prefect", "agent", "start", "kubernetes"]
          imagePullPolicy: "IfNotPresent"
          env:
            - name: PREFECT_API_URL
              value: <http://orion:4200/api>
This what I have for agent
So does it mean that we have to create agent and work-queue with same names & those flows gets picked up
k
Check here. The agent takes in the id of a work queue it starts. It’s not by name. So you need to create a work queue, get the id, and then start an agent. If you click the work queue, it gives you the command to start an agent with it:
k
It seems agent has failed to start
Have followed this tutorial Maybe something strange is going on https://orion-docs.prefect.io/tutorials/kubernetes-flow-runner/
a
you can start your agent with a work queue name provided that you create this work queue (doesn't have to be by ID) looks like issue with network and port forwarding?