Hi all, thanks for a great product! Everything is working great locally, and we've now deployed our meltano project on kubernetes where we already had airflow running. However, there is a hiccup: The job hangs after

Environment 'dev' is active

(see screenshot). After about 15 minutes or so, the actual task completes, but the container keeps hanging telling me that the

meltano run

process has not yet exited and is blocked somehow. To me it seems like it is either stuck in an infinite loop, or trying to call something over the network with (seemingly) infinite retries/timeout. We have a strict firewall policy where we explicitly name the domains to be whitelisted for egress out from our environment (required as we are in the telco business with very strict data protection requirements). My theory is that meltano is trying to call out to some external site. I have disabled telemetry in

meltano

in general (MELTANO_SEND_ANONYMOUS_USAGE_STATS="False" as stated in the docs). Any ideas?

Within the project directory, you should find a file

.meltano/analytics.json

. It contains a dictionary. Set the value of the key

send_anonymous_usage_stats

to

false

.

Can you verify

MELTANO_SEND_ANONYMOUS_USAGE_STATS

is set run

meltano config meltano --list

it'll give you all the meltano settings that should be one. Then you can set ``MELTANO_SEND_ANONYMOUS_USAGE_STATS` either by a bunch of ways a few of which are by

env

variable or in your

meltano.yml

Yes.

https://meltano.slack.com/archives/C01TCRBBJD7/p1659100845294559?thread_ts=1659099832.402809&cid=C01TCRBBJD7 interesting stuff! I would just be worried about new Meltano projects not working due to

.meltano

being ephemeral, but for a test that's a good idea!

If

.meltano/analytics.json

does not exist (e.g. because it's your first run, or you're re-creating the environment each time) then you can build it manually like so (Python code):

import hashlib
import json
import uuid


with open(".meltano/analytics.json", "w") as analytics_json_file:
    json.dump(
        {
            "send_anonymous_usage_stats": False,
            "project_id": str(
                uuid.UUID(hashlib.sha256(b"<your project ID>").hexdigest()[::2])
            ),
            "client_id": str(uuid.uuid4()),  # Any UUID will work for the client ID
        },
        analytics_json_file,
    )

I can do it, but it'll have to wait until monday.

It would be set to false after running a Meltano command. It needs to be set to false before running

meltano

to fully prevent any requests from being sent.

Ah, nice! I'll try that.

An easier alternative to updating

analytics.json

as shown above yourself might be to just run some other command before

meltano run