Hi everyone, my name is Michael. I am happy to be joining the prefect user community. I'd like to report small bug in tutorial. Is this a proper place to do so? Best regards

Back again with more noob questions! I’ve [hopefully] successfully configured my Fargate Agent, to the point where the flow gets picked up and I get the following log output

[2020-06-04 11:01:26,972] INFO - agent | Deploying flow run 05ec3609-4457-4085-8cf3-d01a6d13ccb1

I can see that it’s created the TaskDefinition , but after that it just sits there and never runs the Task, so I’ve no idea what’s going on, and the UI and logs never show any update. Am I missing something? Some magic words to make the agent actually run the task as opposed to just registering? Being able to turn on debug logging would also be a massive bonus 🙂 All I’m getting from the agent is INFO, and I’ve added the

[logging] level="DEBUG"

to the config.toml, and added

export PREFECT__LOGGING__LEVEL="DEBUG"

but nothings happening UPDATE In a fit of desperation I went on to my EC2 box and edited the fargate agent code to make all DEBUG statements INFO, and lo and behold, more log outpu. BUT!!! As far as the

deploy_flow

function is concerned everything succeeded. And I can see the task definition has been created, but nothing else ever happens in log s or the UI. So I’m slightly wiser but not by a lot. 🙂 Even digging into

deploy_and_update_flow_run

in the root agent.py file isn’t giving me much, it seems like everything is succeeding, so I’m at a loss as to why nothing is coming in the UI to show failed, passed, stalled… UPDATE 2 Ok, found something - I took the definition that prefect created and ran it manually, and it gets a

CannotStartContainerError

failed to initialise logging. AccessDenied to create Cloudwatch log group. Fixing. I’ll be back 😂 Does the fargate task need access to call back to the agent/graphql for updates? Finally found some logs and it seems like the flow could be trying to make a callback to localhost:4200, is that right? I realised this was stupid question as I wrote it. UPDATE 3 Turns out it does. From the Fargte Agent docs: Core server

In order to use this agent with Prefect Core's server the server's GraphQL API endpoint must be accessible.

But nothing to say how to configure it? I’ve overridden the cloud.api endpoint in the config.toml and I can see the update coming out in the logs, but the task is still trying to call back to localhost, presumably because the api endpoint override needs to be injected into the task? As in, into the docker container running the flow?

What is the expected behavior if one were to provide a _scheduled_start_time to_ Client.create_flow_run() that had passed/was in the past? https://docs.prefect.io/api/latest/client/client.html#client-2

Is there a better way to dynamically configure which

PrefectSecret

to read from? Our use case is that we have different key/value pairs depending on the lifecycle the Flow is being executed in, so we split that up into two different Secrets, and use a Parameter to trigger which lifecyle it should attempt to process:

with Flow() as flow:
    lifecycle = Parameter("lifecycle", default='dev')
    secrets = PrefectSecret(f"SECRET_NAME_{lifecycle.run().upper()}")

That seems hacky, to have to call

.run()

on the Parameter though. Wondering if anyone found a cleaner approach.

Is there an easy way to alter the

imagePullPolicy

when launching a Job through the Kubernetes Agent? Since the image of the Flow to run will always have a tag for the specific version, this value could be changed from

Always

to

IfNotPresent

. This would decrease the startup time for Flows that get executed frequently, since there's a good chance the image will already exist on the K8 node where it's being executed.

Hi there, another person evaluating Prefect here (it's going very well, btw). I saw https://github.com/PrefectHQ/prefect/pull/2646, which is great since it directly addresses one of my use-cases as well. I'm curious to know if that would also work in conjunction with the

LOOP

construct. The scenario I have in mind is that I have one task that is generating a series of chunks of data. Each chunk would ideally start to be processed by a different downstream task before the generator task completes. I don't know how many it will create in advance, so I don't think I can

map

the task. (It's also possible that I'm just thinking about the problem the wrong way.)

How can i use prefect server to access google cloud?

Trying to upgrade from 0.10.7 to 0.11 -- worked through a few issues, but have been banging my head for awhile trying to figure out what I'm doing wrong here: Flow Runs with newly registered flow version fail immediately with Failed to load and execute Flow's environment: AttributeError("'KubernetesJobEnvironment' object has no attribute 'unique_job_name'"). On closer inspection, Cloud UI still shows Prefect Core Version: 0.10.7 for the flow. I'm using new docker base image FROM prefecthq/prefect:0.11.5-python3.8 and updated

python_dependencies=[
                 "kubernetes",
                 "docker",
-                "prefect[gcp]",
-                "prefect[kubernetes]",
+                "prefect[gcp]~=0.11",
+                "prefect[kubernetes]~=0.11",

(this may be redundant?). I can see in the output from the register/build process:

Step 1/36 : FROM prefecthq/prefect:0.11.5-python3.8
...
Step 30/36 : RUN pip show prefect || pip install git+<https://github.com/PrefectHQ/prefect.git@0.10.7#egg=prefect[kubernetes]>
 ---> Running in e3773eba79d4
Name: prefect
Version: 0.11.5
...

Additionally I have pulled and run the docker image from the registered flow and inside:

root@b3fc72e3e101:/# python
Python 3.8.3 (default, May 20 2020, 20:38:46) 
[GCC 8.3.0] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import prefect
pr>>> prefect.__version__
'0.11.5'

I've now even deleted the flow, all past versions, and registered again and still seeing Prefect Core Version: 0.10.7 in Cloud UI - what am I missing / doing wrong?

Hi, has anyone created a “Flow as a service” system with Prefect? We don’t plan to charge anything, it’s just a collaboration. We want to provide the code for the flows and the end user just provide 1) Data, 2) Dask cluster IP that end user created/paid for independent from us, 3) Result endpoints and we run the flow. We want to let the user provide the compute backend to not have to deal with compute resource billing, and don’t want to see their proprietary data. How does one go about doing that at a high level? Are there any issues, especially security issues, to be aware of?

Also what is the spectrum of deliverables with this? I was thinking, from least work to most work: I was thinking these, from least involved to most involved: 1. Code as a service: • Option 1: Using a third party like Saturn Cloud. We bring the code, User Company only needs to pay for compute power • Option 2: Prefect Cloud, User configure their backend 2. Platform as a service: • We host something on the cloud using customized Prefect server Open source. User Company brings the data and configure backend 3. SaaS: • Option 1: Making an ami/container that is easy to spin up and plug into the resources • Option 2: a complete customized app deployable within User Company’s tools and cluster Am I correct? Did I miss anything?

Hi! I am registering a flow like this:

schedule = IntervalSchedule(
            start_date=datetime.utcnow() + timedelta(seconds=1),
            interval=timedelta(hours=1),
        )

        flow.schedule = schedule
        flow.register()

and it works, but after 10 intervals, it simply stops. Where am I wrong?

Hi everyone - I am trying to understand what

Result.validate

does - and where exactly it is being called (can’t seem to find when its invoked by the runner?) - what does it mean for a task to have an invalid Result ? (any explanation would be much appreciated - thanks)

Hi, newb here trying to find a replacement for airflow. What's the recommended infrastructure to deploy prefect core in production? Looking for minimum CPU/memory. The documentation only mentions a simple pip installation, no real discussion on dimensioning, external dependencies like databases, etc. Is it really that simple? Then why prefect.io cloud?

airflow seemed so complicated to deploy properly for a decent production environment, that I dare not hope

Hi everybody, I have two questions: 1. My Prefect Docker App is called "temp" and all subcontainers subsequently temp_xxxxx. Is there a chance to control the name (instead of having the name by chance)? 2. I need to have all of my flows, tasks, etc, on a local drive. Can I mount that local directory like I do it for other docker container with -v xxx:NAME OF PREFECT-SOURCECODE ? Thanks in advance 🙂 Kai

Hi everyone - We're a Data Science team trying to slowly move towards Data engineering. We've been working on R mostly, and it has served us well, but would like to try and use Prefect for the workflow orchestration. We've a lot of code in R with all the domain logic/tests and so rewriting everything in python is not a viable option. Can we use Prefect for orchestration in this case? I've tried to run the R code as subprocess, and it works to an extent, but it seems very hacky. Is there a standard way to do this?

hi there - I was wondering if there is any support for inter-flow dependencies - as in “run flow B when flow A finishes” - I’m still just getting started, so I might be thinking about this incorrectly - thanks!

@Chris White - shamelessly bugging you because I know you’re there 🙂 - I noticed a PR and commit to remove nomad support. I think the initial support was mostly just a stub, so i get it. Would you be open to PRs for a nomad agent?

Hello, I was wondering what the recommended set up was for running the prefer core server unattended on a virtual machine. The

prefect server

CLI does not seem to support a daemonized mode. Is there some standard way to use the docker-compose config to run the core server using something like Docker Swarm?

Hi, I am looking into getting started with Prefect Cloud. Trying to authenticate my machine with `

prefect cloud login --token $TOKEN

results in

[WinError 10061] No connection could be made because the target machine actively refused it'))

More precisely other prefect auth commands yield the same result. Did I miss anything in the docs in order to establish a connection to prefect cloud?

Hi, is it possible to add a schedule with a default inactive state? I would like to set the Schedule automatically to inactive when a specific env variable is True.

Hi, simple question: did anyone try to run prefect open source components on production? What was the challanges? Really appreciate help here

Hello again, I am running Core/Server at 0.11.5 and have checkpointing set to False in my backend.toml file. Still when the flows run, they create a lot of data in the results-folder which is also not removed automatically. Where else do I have to turn off something to not have this files created, or at least removed afterwards. I suspect that this behaviour is also the reason, why my DaskWorker holds on to a lot of memory after a flow run and then breaks at one point as it does not free up memory at all.

Hi all – I’m just getting started kicking the tires but I’m loving the API so far. Newbie question: What’s the recommended way to handle port conflicts if I’m running a local Postgres server on 5432 on my dev machine? Point Prefect at my local Postgres instance instead of the dockerized service? Use a different forwarded port on the docker service? Something else?

Is there a way to configure automatic flow retries? I am not aware of one, but just wanted to make sure

Hello everyone -- I was trying to find out why a flow I was running paused for over a day and would welcome any troubleshooting suggestions 🙂 I had started a long-running flow with Prefect Core running from an Anaconda prompt on Windows 10. It runs on an AWS VM and basically calls stored procedures on a Postgres RDS database. I kicked it off processing Task A on Thursday saw that it had reached Task B by Friday. It completed Task B on Saturday at 5:22 am and then just hung afterwards. When I looked on Monday AM I saw that the backend had finished Task B but nothing was happening in the terminal. So I went to cancel the flow with a Ctrl-C. But as soon as I did that, it started running Task C after sitting idle for over a day? Here's a snippet from the flow output:

...
[2020-06-05 15:50:12,454] INFO - prefect.TaskRunner | Task 'task a': Starting task run...
[2020-06-05 17:15:04,576] INFO - prefect.TaskRunner | Task 'task a': finished task run for task with final state: 'Success'
[2020-06-05 17:15:04,576] INFO - prefect.TaskRunner | Task 'task b': Starting task run...
[2020-06-06 05:22:27,199] INFO - prefect.TaskRunner | Task 'task b': finished task run for task with final state: 'Success'
[2020-06-08 14:23:12,015] INFO - prefect.TaskRunner | Task 'task c': Starting task run...

Is there anything that I could have accidentally entered in the terminal that would have put it into some type of 'wait' state? Or any other suggestions about things I could check?

Hey, what would you say is the “prefect” way to implement a simple groupBy (or some other reduce operation)? Suppose I have data from several sources (for example partitions) and would like to perform a groupBy+count on the data. With Dask I’d simply load it to a dask dataframe and do the operation on the dataframe. With prefect, the only example shows how to do a simple sum and you basically implement the map-reduce at a low level instead of specifying it logically. So, how would you recommend doing this?

Another question! I notice that some of the tasks on my flow are running multiple times, even though they have finished successfully (so they finish successfully 5 times). What could cause such a behavior?

Hi Community, I am trying out the Prefect Fargate Agent. So far I managed to build Flows with a Docker Storage type, schedule them on Prefect Cloud and launch a Prefect Fargate Agent that deploys the Flows after connecting to Prefect Cloud. Now inside AWS ECS I see that the tasks fail due to:

STOPPED (CannotPullContainerError: Error response from daem)

To my understanding the launched Fargate Instance does not have any access to the Docker Image of the Flow. Concerning the solution from https://docs.prefect.io/orchestration/execution/storage_options.html#docker # Non-Docker Storage for Containerized Environments for rapid deployment, which image would the

metadata={"image": "repo/name:tag"}

refer to in the example given?

Hey guys, I’m trying to run a flow on a custom base_image - it’s built from

prefecthq/prefect:python3.7

and installs a bunch of other junk in there too. The Flow builds and registers, but when I kick it off I get the following:

Failed to set task state with error: ConnectionError(MaxRetryError("HTTPConnectionPool(host='host.docker.internal', port=4200): Max retries exceeded with url: /graphql/alpha

Searching in this channel I see an issue that was identified around this and fixed in April, so I’m not sure if I should still be seeing it, or if there’s config that needs to be passed to the Docker agent to run it? It’s running on Docker agent locally on Macbook if that’s any help.. Prefect version installed locally is 0.11.1, [upgrading now to test] and python in local and the docker image is 3.7 UPDATE: Never mind, upgrade to 0.11.5 did it!