#Retry and back off strategy for DataUpdateCoordinator
jade glacier
Retry and back off strategy for DataUpdateCoordinator
Imho, this is a client library concern
The reasons and logic to retry, the backoff strategy are vastly different across different implementations.
rocky trail
I agree, especially if we follow the separation of concerns principle. My rationale being in the vast ecosystem of HA Core, not all libraries are implmenting it, nor do integration owners always have the ability to influence their upstream library. So basically add it as configurable helper, if you will and add a safety net of transient failures.
Do you see any (architectural) possibilities for a common ground?
delicate snow
What is pretty standard are the retry-after response headers though. Nothing complex, the remote dictates the back-off time.
rocky trail
What is pretty standard are the retry-after response headers though. Nothing com...
That'd be true.