Francesc Gil

While the node is in downtime many things can happen to the cluster, if an object is deleted and the node down has it we should store the action in...

This is an edge case in which on the Replication loop we ask the nodes that do not have the object if they have it before replicating. In the case...

The current implementation uses HTTP by default, but would be good to investigate if switching the transport to GRPC would provide any improvement and if it makes sense to support...

We are just returning errors all over the place but we are not wrapping them, we should use the `%w` to give context to the errors and have the errors...

This should be a background process that would try to balance the storage used between all the Nodes, basically to avoid having situations where 5 nodes are at 90% and...

This is more of a placeholder as I don't know exactly if anything is needed/required on that front. For Policies IDK what could be done/needed. For Security, the infrastructure itself...

We should have a set of metrics to display on the Dashboard: * Total/Node number of Objects * Total/Node number of Replicas? * Total/Node Storage * Total/Node Used Storage *...

If the disk or the [max size](https://github.com/xescugc/rebost/issues/33) is reached then we should not store more objects, but the Node should still be fully functional, for this reason the next Create...

Right now there is no central place that knows where all the objects are and if an object is unique in the cluster, meaning that the same object could be...

If the Node receiving the request of an Object does not have it, it asks the other Nodes if they have it until one answers with the object. In big...