[META] Streaming Model/Agent prediction/execution APIs

[jngz-es]

Is your feature request related to a problem?
The meta issue to track the streaming feature progress.

What solution would you like?

• Model predict rest api based on transport-reactor-netty4 plugin.
• Agent execute rest api based on transport-reactor-netty4 plugin.
• Remote model streaming predict api by supporting sse (sever-sent events) connection with LLMs.
• MLModelTool streaming run api.
• Agent streaming execute api with streaming supported tools.
• Agent streaming execute api with non-streaming tools.

What alternatives have you considered?
A clear and concise description of any alternative solutions or features you've considered.

Do you have any additional context?
Add any other context or screenshots about the feature request here.