MySql check no longer uses Ping

[bgrainger]

What would you like to be added:

Before #1835, MySqlHealthCheck used PingAsync to check the server health. It now defaults to SELECT 1;.

I would like to be able to opt in to using a ping, or even having that restored as the default behavior.

I propose setting the default MySqlHealthCheckOptions.CommandText value to null, and having MySqlHealthCheck interpret that as meaning "ping the server". If CommandText is non-null, then it would be used to form a MySqlCommand.

Why is this needed:

A COM_PING packet will be much more lightweight to send to the server instead of executing a query, even a simple one.

[unaizorrilla]

A PR is welcome on this!

[sebastienros]

There is actually another reason to restore the PING command, with this issue

What is happening is that the Docker image for MySql will start in a temporary state then restart in its final state. The health check based on connection string will report healthy during this pre-initialization state. the binary communication however is not enabled until the final restart is done.

In Aspire I am seeing that the functional tests are failing as they are waiting for the health checks to be successful, but then the mysql calls fail because the server restarts.

[bgrainger]

I'm not sure that using "ping" will help this scenario. If the health check that sends COM_QUERY with a SELECT 1; payload succeeds, then surely a health check that sends COM_PING would also succeed against the first instance of mysqld that's started in a mysql Docker container?

[bgrainger]

@sungam3r Why is this issue marked "up for grabs"? I already have a PR that will fix it: #2093 and included in #2096.