There's very little visibility into the aggregate behavior and operation of logical initial syncs today it's very difficult to answer across a large number of clusters:

• What is the normalized throughput of a logical initial sync
• What is the success rate of logical initial syncs 
• What is the typical duration of an initial sync
• What phase of logical initial syncs do most failures occur
• How many logical initial syncs are there on a given day

To assist in answering these questions and others we should capture relevant metrics. We can take inspiration from resharding metrics.

When complete it should be possible to build a funnel chart/diagram detailing clusters progress through the end-to-end logical initial sync process and charts detailing the performance of logical initial syncs.  

These metrics should survive the end of initial sync, unlike today.