Support for dynamic chunk size

[morgo]

A Pull Request should be associated with an Issue.

We wish to have discussions in Issues. A single issue may be targeted by multiple PRs.
If you're offering a new feature or fixing anything, we'd like to know beforehand in Issues,
and potentially we'll be able to point development in a particular direction.

Related issue: #1204

Further notes in https://github.com/github/gh-ost/blob/master/.github/CONTRIBUTING.md
Thank you! We are open to PRs, but please understand if for technical reasons we are unable to accept each and any PR

Description

This PR dynamically adjusts the chunkSize to be based on execution time feedback. It is disabled by default for backward compatibility, but should be considered a default for future releases. As well as improving performance (my motivation) it is a little bit safer than a static chunkSize considering wide-tables with a lot of secondary indexes.

There are also some guard rails in place to make sure if the observed execution time is 5x the target, it quickly scales down the chunkSize (which is not a problem, as it will later scale back up if this was a temporary blip).

Here's what I observed with some sample tables:

Test	Default Chunk Size (1000)	Dynamic Chunk Size (default 50ms target)
Typical Table (stock)	1m10s(copy)	58s(copy)
Skinny Table (skinnytable)	38s(copy)	30s(copy)

In case this PR introduced Go code changes:

• contributed code is using same conventions as original code
• script/cibuild returns with no formatting errors, build errors or unit test errors.

[meiji163]

This looks promising, I'll do some testing on larger tables.