long-term effects of using python-logging-loki library

[Huda Yousef]

i'm headnig a large project to instrument our tools for logging.

I''ve discussed with my team using this python library to export logs directly to Loki but i've been getting a lot of heat over this decision as the library's latest update is since 2019 which as we all know it not a good library to use. Another argument made against this decision was stating that Loki logic itself might change in the future and that would cause our entire logs system to crash because the library might not be compatible with the new changes.

The alternative option is using backends like Fluentd to scrape container logs, and yes while this is a good option and would offload all log collection from our apps, i really do not like that solution as it complicates a very simple matter we are trying to achieve which is setting a few labels and visualizing our logs. I don't like having to add the complexity of editing config files and deploying backends when i could simply just directly export my logs and creating labels is as simple as writing a dict key/value pair. And out infrastructure is more than capable of handling sending those logs.

Could you please speak with or against such decisions? ... Are their any downsides to using the python library mentioned? ... the library has an MIT license so my team is perfectly able to maintain it however i definitely cannot speak on whether or not Loki-logic changing in the future, which would render that library unusable.

I would really appreciate some feedback on this matter. Is the python library safe to use?