Is there any way to disable color in the error logs? Or maybe just do the normal Python error logging rather than the fancy version? That information is unreadable in my production logs anyway and it would be easier to figure out what’s wrong with simpler error logs.

I think so but I don't know how https://docs.meltano.com/guide/logging#logging If you spend the time and get a good

logging.yaml

for this please share. I honestly dislike the colors and formatting even in my dev environment

I’ve got one of those that turns off all coloring. However, when there are errors, I still get the coloring codes

if you're running in DEV mode Meltano will, but by default that doesn't happen from Meltano

I’ll make a note on https://github.com/meltano/meltano/issues/6482 about your request @sterling_paramore but a thumbsup on the issue would be great too 🙂

What specifically is DEV mode?

I meant DEBUG mode sorry

Nope, I’m only doing logging at INFO level. I don’t think that the log level affects how errors are reported to the log

I haven't experienced that myself

@florian.hines will you chime in here when you get a free moment?

Meltano cli also spits out output that by passes logging completely and sometimes writes colored output (like when theres an error). I'm guessing thats what you're seeing.

Nope: https://github.com/pallets/click/issues/1498

We force a color more often than i realized tbh - https://github.com/meltano/meltano/search?q=fg%3D

Yup, but setting the

color

attribute of the click context to

False

should prevent even those from printing colored output

Ok, so it looks like meltano is using the rich library for error formatting, which does respect the

NO_COLOR

variable.

@ken_payne is working on supporting toggleable pretty printing of exceptions which leverages

rich

. But I'm guessing what you're see'ing is just colorized click.secho output.

🤦 Just realized that yes,

NO_COLOR

does indeed disable color in the error logs. However, it still uses some kind of grayscale for highlighting and thus does not remove the annoying color codes!!!

We were just discussing this and making an action plan. My understanding is that

NO_COLOR

should remove color codes. Admittedly there's still more to do. I've added a comment to also check on greyscale codes but please chime in here with comments/suggestions: https://github.com/meltano/meltano/issues/6482#issuecomment-1234671360

@sterling_paramore what version of Meltano is your screenshot of? As Florian mentioned, I am working on a PR that adds rich as a dependency (I believe for the first time), but it hasn't merged yet 🤔 Is there any chance you have rich in your environment alongside Meltano? Much of my work has been focused on turning rich off in all the relevant places when it isn't needed, specifically because structlog defaults to rich if it finds it installed 🤦‍♂️

Using

meltano==2.5.0

.

rich==12.51

was installed as dependency of

typer==0.6.1

as a dependency of

dagster-cloud==1.0.6

@sterling_paramore - Is it safe to assume from the above that means that

dagster

and

meltano

are both installed on the same python environment or that you are not using python virtual environments? If Meltano is installed with

pipx install meltano

, I would the Meltano bits to be kept separate from other installed libraries - but would be good to confirm.

Yes, they’re both in the same containerized Python environment. I presumed that

pipx

was mostly for developer convenience

Actually

poetry

is the Dev-focused tool but

pipx

is really for all application-type installations.

But I’m used to having a requirements.txt file to pin the versions of all my dependencies.

Yeah, this would work: 1. Install dagster 2. Install pipx 3. Install meltano (with pipx, or manually in a virtualenv) You could also manually create the virtual environment if you prefer to not use

pipx

.- but it's just generally easier to have

pipx

manage the virtual environments for you.

That worked!

COPY requirements/requirements-eta_wh.txt /requirements-eta_wh.txt
RUN pip install --upgrade pip
RUN pip install --ignore-installed -r /requirements-eta_wh.txt && rm -rf /root/.cache

# This instead of `pipx ensurepath`
ENV PATH=$PATH:/root/.local/bin

# Install meltano in a separate python environment via recommended pipx
RUN pipx install meltano==2.5.0

No more extremely verbose fancy colored error logging!

🙌 🎉

Thanks for your help!

Np. This was a tricky one. Glad it got worked out!