yeti.conf api key handling?

[mentalburden]

I use docker compose to spawn yeti and it works great, thanks for the rad project. However, there doesnt seem to be a clean way to handle api keys without tearing down the entire stack or restarting containers, which isnt great for my use case. Ive got a lot of custom feeds that fire observables into yeti through the api, so when the containers go down it can be a noisy event. I know that things like the AlienVaultOTX key is put into yeti.conf, and i can sed/replace a key into that field. But is there a way to refresh the api key listing while the yeti-api container is running instead?

Im picturing the following in my head and it seems feasible to me:
-spin up the yeti docker compose stack
-docker exec -it into the yeti-api container, apt-get update and install nano (or just use vi) to edit yeti.conf to include the api keys you need
-run a py script or a docker compose run command to refresh the yeti.conf file inside of the yeti-api container

Another option would just to include a redacted password field in the frontend so we could copy/paste our key into the ui? This could be something added to the Feeds page next to each of the feeds that requires an api key? Or maybe a section in the System/Staus page that exposes the yeti.conf file and allows for ad hoc edits while the service is running?

Sorry to complain, but its kind of a hassle to have to completely restart a container just to edit/add api tokens for the feeds. Please let me know if theres something else I could be doing instead, or if theres a quick-and-hacky way for me to update the api keys while the yeti-api container and service is running.

[tomchop]

Hey, sorry for letting this hang for so long. I'm trying to understand your need a bit better. Do you actually need the config to persist across container restarts? You can set configuration values through environment variables as well. Have a look at https://github.com/yeti-platform/yeti-docker/blob/main/prod/.env,

The YETI_BLOOM_BLOOMCHECK_ENDPOINT env variable is accessed using:

yeti_config.get("bloom", "bloomcheck_endpoint")

You could also have a YETI_FEEDS_ALIENVAULTOTX=password env variable that you access doing

yeti_config.get("feeds", "ALIENVAULTOTX")  # or lowercase

Does this answer your question? You could also tweak the docker compose file to mount the yeti.conf file on your local filesystem for quicker editing, but the env vars are probably the easiest way to get to what you want (and accessing the setting through python is done in the same way)