Closed dotsdl closed 10 months ago
IP networks are not 100% reliable, so some fraction of HTTP requests from the PortalClient can fail during normal usage. Adding built in use of retries with exponential backoff (e.g. via retry) would smooth over these failures for users/compute.
PortalClient
This would partially address issues raised in #740.
IP networks are not 100% reliable, so some fraction of HTTP requests from the
PortalClient
can fail during normal usage. Adding built in use of retries with exponential backoff (e.g. via retry) would smooth over these failures for users/compute.