Expose go-elasticsearch BulkIndexer output performance metrics

[cmacknz]

With the switch to the go-elasticsearch client we should expose the Elasticsearch client, transport, and BulkIndexer metrics as part of the shipper metrics.

We should capture and expose at least the following metrics as part of the shipper HTTP stats endpoint:

• https://pkg.go.dev/github.com/elastic/go-elasticsearch/v8#BaseClient.Metrics
• https://pkg.go.dev/github.com/elastic/go-elasticsearch/v8/esutil#BulkIndexerStats

We should also ideally expose metrics for:

• The number of times a BulkIndexer flushed because the byte threshold was hit.
• The number of times a BulkIndexer flushed because the timeout expired.

These two metrics may give us a good idea of if additional workers or indexers are needed. Frequently flushing because the byte threshold is hit may indicate the shipper could benefit from additional workers.

• Relates [Design] Evaluate beats observer pattern for output metrics #163