[CCR] Fetches/indexing/writes/reads/bulks in status fields vs readers in task params

[bleskes]

The status reports for the following tasks often refer to fetches (number_of_successful_fetches,number_of_failed_fetches, total_fetch_time_millis,fetch_exceptions). We also have a similar term (which is identical in meaning) call reads (number_of_concurrent_reads). We also have the generic operations_received and total_transferred_bytes and last_requested_seq_no.

On the write sides we have bulk (number_of_successful_bulk_operations, number_of_failed_bulk_operations) and writes (number_of_concurrent_writes, number_of_queued_writes <-- note the different semantics here).

I think we should align our terminologies across the various places. I personally like the read versus write and bulk vs ops. Howabout:

• number_of_succesful_bulk_reads
• number_of_failed_bulk_reads
• total_bulk_read_time
• number_of_ongoing_bulk_reads
• total_operations_read
• total_bytes_read
• last_requested_seq_no_for_reads
• number_of_successful_bulk_writes
• number_of_failed_bulk_writes
• number_of_ongoing_bulk_writes
• number_of_queued_ops_to_write (not too happy with this one)

/cc @jasontedor @martijnvg

[elasticmachine]

Pinging @elastic/es-distributed

[bleskes]

@chrisronline FYI. What ever ends up being deciding here will need a quick update on the UI side.

[bleskes]

@jasontedor @martijnvg and me iterated on this and came up with the following namings. I mark this as a blocker for 6.5.0:

Stats

remote_cluster
leader_index
follower_index
shard_id
leader_global_checkpoint
leader_max_seq_no
follower_global_checkpoint
follower_max_seq_no
last_requested_seq_no
outstanding_read_requests
outstanding_write_requests
write_buffer_size(_in_bytes)
write_buffer_operation_count
follower_mapping_version
total_read_time_millis
total_read_remote_exec_time_millis
successful_read_requests
failed_read_requests
operations_read
bytes_read
total_index_time_millis
successful_write_requests
failed_write_requests
operations_written
read_exceptions
time_since_last_read_millis

Params

remote_cluster
leader_index
max_read_request_operation_count
max_read_request_size
max_write_request_operation_count
max_write_request_size
max_outstanding_read_requests
max_outstanding_write_requests
max_write_buffer_size
max_write_buffer_count
max_retry_delay
read_poll_timeout

[chrisronline]

Looks good.

I'm going to list the fields we are using in the UI and my best guess as to the new field name:

• fetch_exceptions -> read_exceptions
• time_since_last_fetch_millis -> time_since_last_read_millis
• number_of_operations_indexed -> operations_written
• number_of_failed_fetches -> failed_read_requests

The rest seem unchanged.

Is this accurate on terms of the renames?