The SnapGPT Assistant panel is accessible from the SnapGPT button in the Monitor header on every page in Monitor. When you open the panel for the first time, a welcome screen introduces the available skills and important usage guidelines. After dismissing the welcome screen, the panel displays a How SnapGPT can help home screen with the skills relevant to your current page.

Welcome screen

The first time you open the SnapGPT Assistant in Monitor, the Welcome to SnapGPT for Observability! screen appears. It lists the available skills and highlights two important guidelines:

• SnapGPT suggestions should always be verified before you act on them.
• SnapGPT does not request confidential information. Avoid sharing personal details during your interaction.

Click Understood, Let's Begin to dismiss the welcome screen and open the home screen.

Page-aware skills

The home screen shows two sections: Actions available now lists skills you can use immediately on the current page, and Available on other pages lists skills that are inactive on the current page, each with a note and a link to the page where the skill can be used.

The following skills are available in Monitor:

Skills listed under Available on other pages include a clickable link to the page where they can be used.

SnapGPT Assistant panel

The SnapGPT Assistant panel opens docked to the right side of the page. To switch between docked and floating mode, click the panel toggle icon in the panel header (next to the close button):

• Docked mode: The panel is fixed to the right edge of the page. The page content adjusts to sit alongside the panel.
• Floating mode: The panel detaches into a free-floating overlay. You can drag it anywhere on screen and resize it by dragging any edge or corner. The underlying page content is not affected. Use floating mode when you want to view the panel alongside specific areas of the page.

Your dock or float preference is saved and restored in future sessions.

When you navigate away from a page after opening a result, the panel stays mounted and the analysis remains visible. For example, if you click Open in Designer from a suggested fix, Designer opens in a new tab and the analysis stays open in Monitor so you can refer to it while you work.

Data sent to the AI

The SnapGPT Assistant sends operational data from your environment to generate its analysis. Depending on the skill used, this data may include Snaplex telemetry, node alert codes, and pipeline execution error messages and stack traces. The SnapGPT Assistant does not send pipeline business payload to the underlying large language model (LLM). For more information, refer to SnapLogic security and compliance.

Analysis caching

To reduce cost and latency, the SnapGPT Assistant caches analysis results for up to one hour. If you run the same analysis within that window, the panel returns the cached result immediately without generating a new request to the LLM. To force a fresh analysis, wait for the cache to expire or change the input parameters, such as selecting a different time range or Snaplex.

Known issues

• You might experience latency or timeout issues depending on the complexity of the request and the capacity of the LLM provider.
• As with all AI based on LLMs, the SnapGPT Assistant can occasionally produce responses that are incomplete or inconsistent. Always verify suggestions before acting on them.
• On-demand AI analysis for individual node alerts requires an active connection to the SnapGPT Assistant service. If the service is unavailable, the panel displays an error and does not produce a result.