Feature Request
Along with the existing micro service based astronomy shop application, I would like to add:
- Agentic version of astronomy shop where the agent accepts Natural Language queries instead of HTTP load from load generator.
- Astronomy shop MCP server (The agent will use appropriate tool(s) to satisfy the user request.)
- A chatbot for humans to send requests to agent.
- A NL load generator analogous to existing load generator which can send NL queries to agent.
Is your feature request related to a problem?
In the age of LLM automations, it would be good to have a test ecosystem where people can test the capabilities of the agents and discover the new challenges arising from this shift.
The advantage of converting an existing micro service system is that we will get the ground truth traces/logs for a similar agentic execution.
Describe the solution you'd like:
Implementation of the agent and tools for astronomy shop.
Agent
a) Users can provide a yaml configuration for a simple Agent (system prompt + tools) which can be extended to a MultiAgent configuration. The advantage is that users can test out and verify the different architectural advantages. (Eg: Splitting tools across agents might help in reducing the token usage)
b) A static single agent application
Tools
a) Astronomy Shop MCP server - Advantage is easy to use server which is ready made to use. Possible disadvantage is lack of flexibility to modify tool descriptions.
b) User can provide the tools as a config file and we can build the tools and tool serving(MCP or native langGraph tools) dynamically. Advantage is greater degree of flexibility.
LoadGenerator
a) Manually crafted queries analogous to existing load generator HTTP queries. Toggle feature for agent and existing load.
LLM
a) Configure the agent to use any LLM provider.
b) Configure a VCR caching of LLM responses to reuse previously seen responses. This feature will help new users without LLM's with an easy onboarding and reduce the token cost.
Describe alternatives you've considered.
Mentioned above
Additional Context
Tip: React with 👍 to help prioritize this issue. Please use comments to provide useful context, avoiding +1 or me too, to help us triage it. Learn more here.
Feature Request
Along with the existing micro service based astronomy shop application, I would like to add:
Is your feature request related to a problem?
In the age of LLM automations, it would be good to have a test ecosystem where people can test the capabilities of the agents and discover the new challenges arising from this shift.
The advantage of converting an existing micro service system is that we will get the ground truth traces/logs for a similar agentic execution.
Describe the solution you'd like:
Implementation of the agent and tools for astronomy shop.
Agent
a) Users can provide a yaml configuration for a simple Agent (system prompt + tools) which can be extended to a MultiAgent configuration. The advantage is that users can test out and verify the different architectural advantages. (Eg: Splitting tools across agents might help in reducing the token usage)
b) A static single agent application
Tools
a) Astronomy Shop MCP server - Advantage is easy to use server which is ready made to use. Possible disadvantage is lack of flexibility to modify tool descriptions.
b) User can provide the tools as a config file and we can build the tools and tool serving(MCP or native langGraph tools) dynamically. Advantage is greater degree of flexibility.
LoadGenerator
a) Manually crafted queries analogous to existing load generator HTTP queries. Toggle feature for agent and existing load.
LLM
a) Configure the agent to use any LLM provider.
b) Configure a VCR caching of LLM responses to reuse previously seen responses. This feature will help new users without LLM's with an easy onboarding and reduce the token cost.
Describe alternatives you've considered.
Mentioned above
Additional Context
Tip: React with 👍 to help prioritize this issue. Please use comments to provide useful context, avoiding
+1orme too, to help us triage it. Learn more here.