Introduction
This AI Agent is designed to help you quickly find information about ZEDEDA and answer technical questions. This article will help you get the most accurate and helpful responses.
Prerequisites
- This Early Access feature is enabled through feature flags to select customers.
- CSadmin level access is required to enable and disable feature flags.
Enable the AI Agent
The AI Agent is enabled through feature flags. There are two levels of control to enable the feature flags:
- Root - This is the base feature configuration. If the feature is disabled here, it's disabled across all child enterprises.
- Child - This for the specific enterprise. If root is enabled, only then the child enterprise is allowed to enable/disable the feature. The AI Agent can be enabled in a child enterprise, only if it is enabled in the parent enterprise.
To enable the feature on ZEDEDA Cloud, the ZEDEDA AI Assistant (Beta) feature flag is available as part of entitlements. Contact your sales team if it's not part of your entitlement. You need to login as a CSadmin > go to the Enterprise > go to Entitlements > enable the flag. After it is enabled, you can log into the enterprise and see the feature.
Quick Tips
- Be specific - Include system names, IDs, or exact terminology.
- See the Glossary for terminology.
- Ask follow-up questions - The agent remembers your conversation context.
- Try rephrasing - If you don't get a good answer, rephrase and try again.
What the Agent Can Do
The AI Agent is currently equipped to handle multiple types of questions:
Operational Questions
Questions about your specific system's status and real-time information.
Examples:
- "How many of my edge nodes are online?"
- "Show me all edge nodes that have been offline for more than 24 hours."
- "Can you tell me about project X?"
Documentation Questions
General questions about ZEDEDA concepts, features, and how things work.
Examples:
- "What is an edge node?"
- "Explain the types of edge applications.”
- "How does onboarding an edge node work in ZEDEDA?"
API Documentation Questions
Questions about API endpoints, parameters, and implementation.
Examples:
- "What API endpoint would I use to delete an edge node?"
- "Show me the API call to create a project."
- "What parameters does the edge node update API accept?"
Important Limitations
- The agent will only answer questions related to ZEDEDA.
- The agent cannot answer pricing inquiries.
- For pricing questions, please contact: support@zededa.com
Best Practices for Asking Questions
Be Specific
The more details you provide, the better the response will be. There are also some output size constraints, so please ask follow up questions.
| Avoid | Prefer |
| "How does that work?" | "How does checking an edge node’s status work?” |
| "Show me the events" | "Have there been any critical events in the last 24 hours?" |
| "What's the API for apps?" | "What's the API endpoint to update edge applications?" |
Use Helpful Phrases
If your query isn't generating the response you need, try prepending your query with the following phrases:
| Question Type | Phrase |
| Documentation | “Consult the documentation and explain..." |
| API Documentation | Include the phrase "API endpoint" in your query. |
| Operational / System Status | "Look at my status and …" |
You can also ask for the agent response to be displayed in a graph (i.e., bar chart, pie chart) so that the information is more easily digestible and visually appealing, allowing for quicker insights and better understanding of complex data.
Ask Follow-Up Questions
The agent maintains conversational context, so you can ask for clarification and any follow-up questions you may have. You can also switch between topics within the same conversation.
Understanding Agent Responses
What to Expect
- The agent will often tell you when it's consulting documentation or calling a tool to gather data
- You can request specific actions: "Please check the documentation for..."
- Responses may include code snippets, API examples, or step-by-step instructions
Response Time
- Complex operational queries may take longer as the agent retrieves system data
- You'll see an indicator when the agent is generating a response
Troubleshooting
If You Don't Get a Response
- Click the New Chat icon in the upper right corner.
- Rephrase your question to be more specific.
- Try adding context or using the helpful phrases previously mentioned.
If the Response Isn't Clear
- Ask follow-up questions in the same chat.
- Try rephrasing with more specific terminology.
- Request examples or step-by-step explanations.
Common Error Messages
- “I am currently unable to process your request due to resource constraints. Please try again later."
- Open a new chat window and try your query again.
- Error Code 529.
- This means too many requests on the provider's side.
- Wait a few minutes and try your request again.
- No response generated by the agent and the agent appears unresponsive.
- Open a new chat and try your query again.
Note: When the agent is generating an answer, avoid clicking the following:
- New Chat
- Chat history
- Other options in the chat window
Clicking these will stop the response generation. If this happens accidentally, open a new chat and ask your question again.
When to Contact Support
Contact our support team at support@zededa.com if:
- You need pricing information.
- Your question is outside the scope of the present documentation.
- You've tried rephrasing multiple times without success.
- You need urgent assistance with a critical system issue.
Feedback
Your feedback helps us improve the agent. Use the thumbs up and down buttons to rate responses, and click the feedback icon to provide additional details.
Getting Started - Try These Examples
Ready to begin? Here are some sample queries to try:
- “What is a datastore and how can I create one?”
- “How do I onboard an edge node?"
- “Give me an overview of my edge app instances."
- "What API endpoint should I call to create a project?"