Satisfy ResourceManager debug needs with smaller code footprint

[BigLep]

Done Criteria

It is possible for a user to follow docs at https://github.com/ipfs/kubo/blob/master/docs/libp2p-resource-management.md to understand:

1. What the limits are for a running Kubo node
2. What the current utilization is against those limits
3. What limits are closest to being fully utilized

Why Important

We have a lot of code right now in our stat/limit commands for printing and analyzing limits. The problems with it are:

1. Makes keeping up-to-date with libp2p refactors (like what's happening in WIP: Implement new ResourceManager partial limits config object. #9612) more difficult
2. We aren't flexible in highlighting the information that we want (e.g., most important scopes first, limits that are most highly utilized only)

Notes

The current proposal is to have two commands from Kubo:

1. Get the limits of a running Kubo node. We should ideally get the root object from the go-libp2p resource manager/accountant and print it out as JSON
2. Get the current accounting/utilization of a running Kubo node. We should ideally get the root object from the go-libp2p resource manager/accountant and print it out as JSON.

With these sets of information, it's possible to run jq/awk/etc. commands to give output like:

Scope	Limit Name	Limit Value	Current Resource Accounting	Utilization Percentage
System	ConnsInbound	100	50	50%
System	ConnsOutbound	200	10	5%
Transient	ConnsInbound	10	4	40%
Transient	ConnsOutbound	50	10	20%

This is really the key summary users or maintainers need for helping debug.

This assumes we're also doing the work of removing Swarm.ResourceMgr.Limits in favor of limits.json in go-libp2p: #9603 . With that, we are free from the concern of our commands printing out values that can be copy/pasted into Kubo config.

Doing this work should allow us to drop other things on the backlog:

• Improve the utility of the swarm stats command for determining which limits are being hit #9474
• Resource Manager: ipfs swarm limit <scope> --reset is setting to zero all other scopes. #9559
• ipfs swarm limit all --reset does not remove user overrides from Swarm.ResourceMgr.Limits #9576

In addition to code changes, this work will involve:

1. Updates to https://github.com/ipfs/kubo/blob/master/docs/libp2p-resource-management.md
2. Updates to https://github.com/ipfs/kubo/blob/master/docs/config.md#swarmresourcemgr
3. Ensuring we longer have the problems listed in ipfs swarm limit all: returns empty "" Service and Protocol name #9577

[ajnavarro]

There are some inconsistencies in libp2p stats output and limits output (different fields):
Stats:

ipfs swarm stats | jq .System
{
  "Memory": 13893632,
  "NumConnsInbound": 0,
  "NumConnsOutbound": 1295,
  "NumFD": 63,
  "NumStreamsInbound": 15,
  "NumStreamsOutbound": 186
}

Limits:

ipfs swarm limit | jq .System
{
  "Conns": 1000000000,
  "ConnsInbound": 16212,
  "ConnsOutbound": 1000000000,
  "FD": 524288,
  "Memory": 17000000000,
  "Streams": 1000000000,
  "StreamsInbound": 1000000000,
  "StreamsOutbound": 1000000000
}

Checking if that is intended.

[ajnavarro]

Related PR #9623

I added a new command with the table view because was difficult to create using two JSON outputs.