feat(share/discovery): Discard method

[Wondertan]

Fixes #2204

Regarding the case where we exhausted all the peers on the network and backed them all off. It can happen on the smaller network, and in the worst case, the node will wait 10 minutes(the default backoff time), which is acceptable considering peers from shrexsub #2105

[walldiss]

It seems like with current solution peer could be rediscovered right away. Could we move discarded peer to some set that will prevent it from be in discovery set?

[walldiss]

Nvm, answer was in description. I'm a bit sleepy and skipped reading it

[Wondertan]

...

=== RUN   TestIntegration/get_peer_from_discovery
    manager_test.go:446: 
        	Error Trace:	/home/runner/work/celestia-node/celestia-node/share/p2p/peers/manager_test.go:446
        	Error:      	Received unexpected error:
        	            	context deadline exceeded
        	Test:       	TestIntegration/get_peer_from_discovery

[distractedm1nd]

For the integration it may be easier if the Discard API returns a boolean, if it was in discovery or not. This is because:

For removing a peer from the pool, we rely on discovery to call our callback, which calls pool.remove
If we call disc.Discard(peer), it will call pool.remove through the callback
But not all peers are from discovery, so if Discard didnt get rid of the peer we need to call pool.remove ourselves

Not ideal but its a solution