[metrics] Add OpenAI metrics about chat completions and text completions

[eolivelli]

Summary:

• add metrics about OpenAI chat and text completions
• add a script to start Prometheus and Grafana for docker run users
• please note that when you are using the Streaming mode OpenAI doesn't respond yet with the usages (see this ticket https://community.openai.com/t/openai-api-get-usage-tokens-in-response-when-set-stream-true/141866/19)
• in the "Flare" example I am disabling streaming output, this way you can see some values on Grafana (also for Flare it is useless to use Streaming output)