This solution removes the panic when there is not enough capacity and discards the data. It flags the data loss internally, then the next time there is enough buffer capacity it sends a message to inform that there was a loss plus the logged message and resets the flag.
I think that ci/upload-docs.sh is failing in Travis because the PR comes from a fork (then no secrets/credentials are available, see this). Something similar happened to me once. Anyway let me know if there is anything else I need to do ;-)
Addressing #69
This solution removes the panic when there is not enough capacity and discards the data. It flags the data loss internally, then the next time there is enough buffer capacity it sends a message to inform that there was a loss plus the logged message and resets the flag.