Non-standard Environment Variable Names

[tebaly]

I would like to raise an issue regarding the naming conventions of environment variables in the Litestream project. Currently, the project suggests using non-standard variable names for authentication with AWS services:

kubectl create secret generic litestream \
    --from-literal=LITESTREAM_ACCESS_KEY_ID="..." \
    --from-literal=LITESTREAM_SECRET_ACCESS_KEY="..."

Instead, it would be more beneficial to align with standard practices and use the following variable names:

AWS_ACCESS_KEY_ID
AWS_SECRET_ACCESS_KEY
AWS_DEFAULT_REGION

Using these standard variable names would allow users to leverage existing secrets already present in the Kubernetes cluster, meant for AWS authentication. However, due to the non-standard naming conventions, users are forced to duplicate the same access credentials with different variable names. This redundancy adds unnecessary complexity and hampers the usability of the Kubernetes cluster.

I kindly request the Litestream project team to consider updating the documentation and examples to promote the use of standardized environment variable names for AWS authentication. This change would enhance compatibility and improve the overall user experience with Litestream in Kubernetes environments.

Thank you for your attention to this matter.

[talmeme]

litestream works with more than just AWS S3. I'm using litestream with Oracle Cloud's S3-compatible storage. It would be confusing to be using AWS_ variables with non-AWS S3-type storage.