System transaction for critical metrics

From a node operator's perspective, it would be convenient if there was a system transaction that carried various metrics from all nodes. Currently, a node operator can only get access to the metrics of another node if it has access to a shared grafana dashboard (or equivalent). In the future, there may be node operators who decide not to share metrics data.

Before creating this feature, we should identify a small subset of mission-critical metrics.

- ping time between neighbors
- bandwidth utilization between neighbors
- transaction queue size
- TODO: list others

There should be configuration that causes this transaction to be submitted by each node every X seconds.

When these transactions are received, each node would unpack them, and then export the data to the metrics engine.

This allows all node operators to see the critical metrics from all honest and healthy nodes (as long as we don't lose quorum), even if those node operators do not decide to export metrics data.

Start by mapping the system transaction submission and receipt-handling paths, then inspect how received data reaches the metrics engine. Define the initial mission-critical metric subset, the configurable submission interval, and the format and export behavior needed for operators to see data from other nodes.

Written by the indexing model from the issue text.