Docs » AI Assistant in Observability Cloud

AI Assistant in Observability Cloud đź”—

The AI Assistant in Observability Cloud provides fast, deep insights that leverage observability data (metrics, traces, logs, and alerts). Users can access these insights in an intuitive chat interface within their existing workflow in Splunk Observability Cloud.

You can use the AI Assistant to do the following:

  • Get quick insights into your metrics, charts, traces, incident alerts, dashboards, services, and errors

  • Troubleshoot faster in Splunk APM and Splunk Infrastructure Monitoring

  • Search metrics, charts, and alerts

  • Generate SignalFlow based on your question in natural language

The AI Assistant provides an intelligent chatbot experience to empower users to craft complex SignalFlow by simply writing plain English prompts.

The AI Assistant has access only to data from Splunk APM and Splunk Infrastructure Monitoring. While you can access the AI Assistant throughout Splunk Observability Cloud, responses are geared toward use cases for APM, Infrastructure Monitoring, and Log Observer Connect.

What is supported đź”—

The AI Assistant understands and supports natural language. The AI Assistant responds in natural language with a summary of insights synthesized from multiple sources. Currently, the AI Assistant supports only English.

How to access the AI Assistant in Observability Cloud đź”—

To access the AI Assistant in Observability Cloud, select the AI Assistant icon from the toolbar on the right side.

This image shows the location of the AI Assistant in Observability Cloud.

Enter your prompt in plain English in the text box. Ask about anything in your environment.

Alternatively, you can select from the suggested prompts:

  • What can you help with?

  • List active incidents

  • List error traces from the past 15 minutes

This image shows the suggested prompts for the AI Assistant.

Note

When you ask questions or submit instructions that require the AI Assistant to query logs, there is an impact on your SVC quota.

Feedback đź”—

You can give feedback to the AI Assistant in development team by selecting the thumbs up or thumbs down icon that appears after each AI Assistant response.

Context limitations đź”—

Context length is the amount of text that the AI Assistant can process in a single conversation due to LLM limitations. Think of it as the short-term memory of the AI Assistant. Each interaction with the AI Assistant is limited to the context set from the beginning of that conversation. The AI Assistant tells you when you exceed the chat limit and prompts you to copy the information that you want to save and start a new chat to continue. Select the New chat button in the upper right corner of the AI Assistant to create a new chat with a new context.

Chat history đź”—

You can only access the most recent chat interaction within the last thirty days.

ChatId đź”—

Chatid is the unique identifier for each conversation in AI Assistant in Observability Cloud. Use ChatId when you want to report something about a particular conversation. Find it near the top of the chat below AI Assistant.

This image shows the location of the chatId.

Data sharing and use đź”—

When you interact with AI Assistant in Observability Cloud, Splunk collects and might use inputs, outputs, grounding data, feedback, and usage data to provide and maintain the the AI Assistant, comply with applicable law, enforce our policies, and develop and improve the AI Assistant in Observability, including to train AI models.

If you do not want your data used for these purposes, do not click on the button and do not install, download, access, or otherwise use AI Assistant in Observability.

The following table explains the categories of data the AI Assistant collects:

Category

Description

User prompts or inputs

Refers to a question or an input by a user to the AI Assistant. Examples are “Show me all K8 nodes with more than 90% memory utilization”, and “What is wrong with my payment service?”

Grounding observability data

Refers to observability metrics, traces, and logs data. Not every user prompt may require grounding observability data. For environment-specific questions like “What is wrong with my payment service?”, the AI Assistant leverages payment service-related observability data to answer the question.

Assistant responses

Refer to the output generated by the AI Assistant. This might contain observability data in a summarized text chart form.

Feedback

Refers to any user-entered feedback.

Usage data

Usage data is more fully described in the Splunk Privacy Statement. Examples include “thumbs up”, “thumbs down”, “chat id”, “copy”, “tokens used”, and “response length”.

You can activate and deactivate data sharing by going to Settings. On the General Organization Settings page, go to the AI Assistant Management section. Select or deselect AI Assistant in Observability to activate or deactivate the AI Assistant for your organization.

If you do not want to share your data, deselect Allow my AI service data to be used for research and development purposes.

This screenshot shows how Splunk Observability Cloud products serve the different layers and processes in an organization's environment.

Service limitations đź”—

A Splunk Observability Cloud organization has a limit of 3,000 prompts per month and no more than 10 prompts per minute.

Other resources đź”—

Learn how to prompt the AI Assistant effectively by reading Prompt guide and library for AI Assistant in Observability Cloud. You can find general guidance on how to prompt the Assistant to achieve the best results, as well as see sample prompts and learn what makes a prompt excellent, good, or poor.

To learn about Splunk’s commitment to responsible AI, see Responsible AI for AI Assistant in Observability Cloud .

This page was last updated on Feb 05, 2025.