Readiness endpoint could check if upload was successful

[agazso]

On the public gateway there is an issue with the Bee nodes that may cause the nodes being full and therefore failing to upload data to them (see ethersphere/bee#3194).

It would be good if the /health endpoint would try to upload a single chunk worth of random data to check if this is the case and if the upload fails then return an error, so that it can be removed from the load balancer.

[vandot]

Good idea just please use /readiness endpoint for it there is already some check under it
https://github.com/ethersphere/gateway-proxy/blob/master/src/server.ts#L78

[agazso]

Good idea just please use /readiness endpoint for it there is already some check under it
https://github.com/ethersphere/gateway-proxy/blob/master/src/server.ts#L78

Thanks, changed the title accordingly.

[Cafe137]

Should we use deferred or direct upload for the single chunk?

[agazso]

Deferred, because the goal is just to check if the node works correctly, not the network itself.