Proposed changes to default storage adapter about pan-cortex-data-lake-python HOT 2 CLOSED

I will be shipping the storage_params suggestion since it would be a non-breaking change to the existing TinyDBStore class.

from pan-cortex-data-lake-python.

Observed differences between credentials.json and MemoryStorage:

• credentials.json acts as a true, shared credentials store. This means clients are (potentially) able to read and write to this shared store from different runtimes/execution environments.
• MemoryStorage mode is not a shared credentials store. This means clients are only able to read and write to this store within their respective runtime/execution environment.

Implications:

• In serverless deployments, which implement distinct execution environments, it will be required to implement a Credentials subclass or custom storage adapter in order to share, reuse or cache credentials across function calls.

from pan-cortex-data-lake-python.