What is the best way to tell Kubernetes Agent to utilise only the nodes from a specific Node pool? We are running the agent on AKS in a separate namespace.

Is it possible to re-run tasks in already running flow? I have quite a big DAG of tasks. If one task failed, its downstream tasks are triggerfailed. But the run still continues because there are many other parallel tasks which are running. So imagine i fixed the issue because of my task failed. How do i re-run it within existing running flow? Or should i wait for flow run to complete and only then restart flow run from failed?

So i was able to narrow it down to a such case: imagine you have a flow with two branches, one branch is a long-running and other is short lived and may fail. So if short branch fails, its dependencies are triggerfailed. If i re-run failed task in a still running flow, its state is set to pending and nothing happens. Flow finishes when long task fininished and this task is still in pending state. But, if before re-running failed task i set downstream triggerfailed task to pending state, re-running failed task executes as expected. Is this a bug?

How do I unpickle a file that I used the

PickleSerializer()

to store during a prefect flow?

The task decorator on the function looked like this:

@task(
    checkpoint=True,
    result=GCSResult(
        bucket="my-bucket",
        location=RESULT_LOCATION_FORMAT,
        serializer=PickleSerializer(),
    )
)

Hi. I am wanna test local debugging based on this document: https://docs.prefect.io/core/advanced_tutorials/local-debugging.html#locally-check-your-flow-s-docker-storage I have created a task like this:

import prefect
import time
from prefect import task, Flow
from prefect.environments.storage import Docker


@task()
def whoami():
    logger = prefect.context.get("logger")
    time.sleep(200)
    <http://logger.info|logger.info>('Hello world')
    return 'hello World'


storage = Docker()
flow = Flow("reddit-flow", storage=storage, tasks=[whoami])
built_storage = flow.storage.build(push=False)
print(built_storage.flows)

and run it

python code.py

and the image successfully created. but the flow doesn't exist inside the image. neither inside

/root/.prefect/

nor

/opt/prefect/flows/

. but when I change the code to register the flow, the created image has the flow. (also i think local debugging documentation is outdated) am I missing something? thank you in advance

Is the only way to delete project using non-cloud, self-hosted prefect via graphql ala https://docs.prefect.io/orchestration/concepts/projects.html#graphql-2 ? Was hoping to do via UI but it looks like only cloud version has that feature?

Would this be a reasonable way to say "Run this TUESDAY through SATURDAY"?

not_filters=[
            filters.is_day_of_week(1),
            filters.is_day_of_week(7)
        ],

Hello I'm experiencing a problem that I don't know how to solve. I have the following program (agent.py) that starts with the supervisor when starting my Linux machine. It aims to load all the folders and files of my project where the flows are.

from prefect.agent.local.agent import LocalAgent
from os.path import dirname, abspath, join
from os import listdir, path as p,walk

path = abspath(join(dirname(__file__)))

import_paths=[]
for file in listdir(path):
    if not file.startswith(".") and p.isdir(file) and file in ['db','service','flows','util'] :
        import_paths.append(join(path,file))

for path_import in import_paths:
    paths = [p.join(root, name)
                for root, dirs, files in walk(path_import)
                for name in dirs
                if not name.startswith(".") and not name.startswith("__")]
    import_paths.extend(paths)

LocalAgent(import_paths=import_paths,show_flow_logs=True).start()

My Supervisor File is as follows:

[program:start_agent_prefect]
command=python3 /home/dev/projetos/prefect-tasks/agent.py
user=dev
autostart=true
autorestart=true
stderr_logfile=/var/log/agent_prefect.err.log
stdout_logfile=/var/log/agent_prefect.out.log

However, the prefect agent is unable to execute the Flows as it cannot find the prefect giving the message below:

[Errno 2] No such file or directory: 'prefect'

[2020-11-20 22: 21: 28,139] INFO - agent | Found 1 flow run (s) to submit for execution.
[2020-11-20 22: 21: 28,176] INFO - agent | Deploying flow run 3ab508f4-731c-450b-8b30-bdaaaeba6117
[2020-11-20 22: 21: 28,180] ERROR - agent | Logging platform error for flow run 3ab508f4-731c-450b-8b30-bdaaaeba6117
[2020-11-20 22: 21: 28,232] ERROR - agent | Error while deploying flow: FileNotFoundError (2, 'No such file or directory')

Note: If you manually run the agent.py file in the folder with: dev $ python3 agent.py The agent works normally, could someone give me a light where I am failing?

Are there any good examples of using signal.SKIP() to short circuit running the whole flow? Trying to wrap our brain around it. Example:

Hi! I'm interested in measuring percentage of tasks in the pipeline that were read from the cache. The only way to obtain that information that I can think of would be to have a shared counter that would be incremented at the beginning of each task. I believe that global variable won't work when running in distributed mode so I was wondering about an alternative. Is it possible to modify prefect.context from within a task? And then access such accumulator in a task that runs at the very end and prints the result/saves it somewhere?

Hi! does anyone know if there is a way to copy a Flow from one project to another programmatically?

using prefect cloud

Hey team, is there a recording of this talk? We are interested in using the helm chart in our deployment

Hi! I'm new here, and new to open source/contributing as well, I'd love to contribute to the project and I've went through the contribution part of the docs, read some good first issues and cloned the repo locally, but I'm still a little lost on how to go about solving any issues. Could someone please help me on getting started?

Hi all! I'm looking at using Prefect to manage ETL processes that are executed by PySpark on a CDSW/Hadoop environment. I'd like to subclass the Results class---much like what has already been done for Google Cloud and AWS---such that each task can read/write to Hive storage by toggling the Prefect cache variable. I.e. loading a cached task would simply read the spark dataframe created by that task via

sql_context.read.table

. Are there any references for this?

If I use a Kubernetes agent to run my flows, what would be the recommended way to execute non-python code? Spawn off new Kubernetes jobs in the cluster, or are there other ideas on this?

I'm registering a flow on a locally hosted server and would like the extra loggers functionality to be in effect when the scheduled flow runs. It looks like setting the environment variable `

PREFECT__LOGGING__EXTRA_LOGGERS

as per https://docs.prefect.io/core/concepts/logging.html#extra-loggers during flow registration isn't sufficient -- I'm not seeing my apps

logging

statements in the web ui. Is there a way to get this env variable setting to be in effect for when the app runs?

Hiya, trying out the

DaskKubernetesEnvironment

and am running into a weird issue.

Hi, I was wondering what people do when coming across parts of objects that are not serializable by Cloudpickle when trying to save some result objects to GCS. The result objects are heavily nested so it is difficult to identify which parts of them are not pickleable. Is it considered poor practice to try to send the whole result object to GCS? Or is there some way around it?

Hi, maybe I missed the relevant documentation, but what exactly does a python module need for the CLI to be able to register a flow? Is it a module-level variable called

flow

?

Hi everyone, I'm using the newly announced k8s helm chart, which is awesome by the way! I'm trying to add basic authentication to the apollo server to protect it through an nginx controller on kubernetes. But now the "server cannot be reached". Could anyone please kindly offer up any insight on this? Would be most appreciated! The other alternative I could look into is making the apollo server an internal load balancer which I'll try next.

Hey guys, I have Prefect Server and Dask Scheduler running on machine A and Dask Worker on machine B (machineB has no Prefect or Docker). When I tried to send a job from machine A to machine B to execute by using the task tag described here:https://docs.prefect.io/api/latest/engine/executors.html#executor, error occurred like this on machineB:

distributed.protocol.pickle - INFO - Failed to deserialize b'\x80\x04\x950\x00\x00\x00\x00\x00\x00\x00\x8c\x1dprefect.engine.executors.dask\x94\x8c\n_maybe_run\x94\x93\x94.'
Traceback (most recent call last):
  File "/anaconda3_a/envs/py37/lib/python3.7/site-packages/distributed/worker.py", line 3268, in loads_function
    result = cache_loads[bytes_object]
  File "/anaconda3_a/envs/py37/lib/python3.7/site-packages/distributed/utils.py", line 1573, in __getitem__
    value = super().__getitem__(key)
  File "/anaconda3_a/envs/py37/lib/python3.7/collections/__init__.py", line 1027, in __getitem__
    raise KeyError(key)
KeyError: b'\x80\x04\x950\x00\x00\x00\x00\x00\x00\x00\x8c\x1dprefect.engine.executors.dask\x94\x8c\n_maybe_run\x94\x93\x94.'
...
ModuleNotFoundError: No module named 'prefect'

I thought that I must install prefect on machine B for the task to run remotely. So then I installed prefect and reran the task, a new error occurred like this on machineB:

[2020-11-24 17:02:05+0800] ERROR - prefect.CloudTaskRunner | Failed to set task state with error: ConnectionError(MaxRetryError("HTTPConnectionPool(host='localhost', port=4200): Max retries exceeded with url: /graphql (Caused by NewConnectionError('<urllib3.connection.HTTPConnection object at 0x7ff6099fcf10>: Failed to establish a new connection: [Errno 111] Connection refused'))"))
Traceback (most recent call last):
  File "/anaconda3_a/envs/py37/lib/python3.7/site-packages/urllib3/connection.py", line 160, in _new_conn
    (self._dns_host, self.port), self.timeout, **extra_kw
  File "/anaconda3_a/envs/py37/lib/python3.7/site-packages/urllib3/util/connection.py", line 84, in create_connection
    raise err
  File "/anaconda3_a/envs/py37/lib/python3.7/site-packages/urllib3/util/connection.py", line 74, in create_connection
    sock.connect(sa)
ConnectionRefusedError: [Errno 111] Connection refused
...
2020-11-24 17:02:05+0800] INFO - prefect.CloudTaskRunner | Task 'test_task': Finished task run for task with final state: 'ClientFailed'
[2020-11-24 09:02:10+0000] CRITICAL - CloudHandler | Failed to write log with error: HTTPConnectionPool(host='localhost', port=4200): Max retries exceeded with url: /graphql (Caused by NewConnectionError('<urllib3.connection.HTTPConnection object at 0x7ff6097590d0>: Failed to establish a new connection: [Errno 111] Connection refused'))
[2020-11-24 09:02:10+0000] CRITICAL - CloudHandler | Failed to write log with error: HTTPConnectionPool(host='localhost', port=4200): Max retries exceeded with url: /graphql (Caused by NewConnectionError('<urllib3.connection.HTTPConnection object at 0x7ff6097590d0>: Failed to establish a new connection: [Errno 111] Connection refused'))
[2020-11-24 09:02:10+0000] CRITICAL - CloudHandler | Failed to write log with error: HTTPConnectionPool(host='localhost', port=4200): Max retries exceeded with url: /graphql (Caused by NewConnectionError('<urllib3.connection.HTTPConnection object at 0x7ff6097590d0>: Failed to establish a new connection: [Errno 111] Connection refused'))
[2020-11-24 09:02:10+0000] CRITICAL - CloudHandler | Unable to write logs to Prefect Cloud
[2020-11-24 09:02:10+0000] CRITICAL - CloudHandler | Unable to write logs to Prefect Cloud
[2020-11-24 09:02:10+0000] CRITICAL - CloudHandler | Unable to write logs to Prefect Cloud

I am really confused by “ConnectionError” and “Unable to write logs to Prefect Cloud”. Do you guys have any idea why this might be happening? Must I install Docker (on machine B) as well for the task to run?

Hey, I'm (yet another) person who would like to have a more flexible logging configuration in Prefect. I was a bit surprised that Prefect does not allow specifying logging via

logging.dictConfig()

-like format, especially since Dask supports it (https://docs.dask.org/en/latest/debugging.html#logs). If you could point me in the right direction where the logging setup happens (I think I found the agent's logging setup, but not sure about task logging, I assume that's separate for each execution environment?), I would be willing to work on a PR to allow specifying logging via dict format. For reference, I have `uvicorn`'s server, which has similar options to Prefect, but also allows passing the entire logging configuration.

Hi, i was wondering what is the best practice to cancel a flow run from one of our task ? Currently i'm raising a

signal.ENDRUN(state.Cancelled())

which put my task in the right state, but the following tasks are still executed meaning the trigger all_successfull is still true. Any idea what am i doing wrong ?

Hi, I have a question about a family of built-in tasks

ReadGsheetRow

,

ReadAirtableRow

etc. What is their use case? All of these tasks Read- and Write- counterparts receive

n

- identifier of the row to be read or written. In our experience there's no such scenario when you need just a single row from table. Even more: there's no way to find out the size of the table. So is anybody actually using these tasks in this specific implementation?

Question about Run Configuration for Flow. What is a meaning behind specifying LocalRun/DockerRun/.. in order to specify list of required labels for agent? What if I don't care what kind of run it will be as long as it satisfies required labels? How do I specify labels without specifying a type of run?

Suppose I have a long sequence of tasks within a given flow, where each task reads some persisted data created by the previous task, operates on that data, and persists the resulting output to a database via checkpointing. How can I... 1. Run a single task within the flow and pull its input data from the persisted checkpoint of the previous task, i.e. debug a single task. 2. Have my flow rebuild the persisted output of any given task, i.e. ignore the fact that the checkpoint exists, recompute and overwrite it. Thanks in advance! Love the package and selling my team hard on it!

For the Prefect Server Helm Chart, is there a specific reason as to why it's not recommended to deploy the packaged postgresql database that comes w/ the helm chart into production? What are the risks involved?