This PR adds a new RPC method called savefeeestimates in a similar fashion to savemempool RPC.
The rational behind this is that currently there is no way to save the fee estimates without shutting down the node and this makes it harder/slower to bring up new instances of bitcoin core.
For example in https://github.com/Blockstream/esplora/ when we do new deployments or when we have higher traffic we need to bring up new instances, currently the procedure uses a snapshot of the blockchain (say a few weeks old) brings it up to the tip, then stops it, calls savemempool on a long running instance, copies that file over and restarts the new node so that the new instance has all the transactions in mempool even though it wasn’t live to receive them when they occurred.
However the fee estimation of that new node are not up to date and as such the new node will report bad fee estimates.
This PR will make it possible to start a new node with reasonable fee estimates from an old running node without having to shut down said old running node.
Any feedback welcome and thanks for reviewing/reading.