[FEATURE] Add query latency metrics to enhance performance monitoring

Is your feature request related to a problem?

Currently, only request count metrics are available in SQL plugin, and there are no query latency metrics. This limitation makes it challenging to pinpoint slow SQL translations and effectively address performance bottlenecks.

What solution would you like?

I propose adding latency metrics to the SQL plugin. These metrics would provide insight into the time taken for SQL translation, helping to quickly identify and resolve performance issues.

What alternatives have you considered?

Without direct latency metrics, we have to rely on ALB latency metrics and OpenSearch search latency metrics as indirect indicators, which is less precise.

Do you have any additional context?

Current available metrics: https://github.com/opensearch-project/sql/blob/main/docs/user/admin/monitoring.rst