mma: consolidate logs in one msg

**Describe the problem**

Ideally, we should emit a single summary message with enough info to understand why an overloaded store isn't shedding ranges away. Something like:

```
s23 load shedding blocked: 15 attempts rejected
no-candidate-to-accept-load[ByteSize]: 12 rejected
e.g. r1 → s23: load=[cpu:568ms/s, byte-size:4.8 MB], capacity=[cpu:899ms/s, byte-size:84 MB]
constraint-violation: 3 rejected
e.g. r5 → s12: violates zone constraint "region=us-east1"
```

This should be at V0 level.

Jira issue: CRDB-59345

Epic CRDB-56265