Monitor Retool Agents
Learn how to monitor Retool Agents for performance, reliability, and cost-management.
The Monitor page for Retool Agents has three main components: a global view showing all agents, a detailed individual agent view, and an individual agent run view.
The key benefits of the Monitor page are:
- Reduce time to identify and resolve agent issues.
- Improve understanding of resource usage patterns.
- Enhance ability to optimize agent workflows.
- Improve visibility into cost drivers.
- Provide actionable insights for runtime optimization.
All agents view
The Monitor page is comprised of several sections that detail information for all agents.
Monitoring page for all agents.
The agent sidebar
The agent sidebar includes an expandable list of all agents, small activity graphs showing recent runs, and a search bar. Click on agent from the sidebar to access monitoring information for that particular agent, or expand an agent and click on a particular run to access the monitoring information for that agent run.
Use the agents sidebar to view monitoring information for all agents, a particular agent, or an agent run.
Time range selector
The time range selector at the top of the page shows runs and errors across all agents for the selected timeframe when hovering over the graph.
The time range selector allows you to view the run and error history of all agents or an individual agent.
- The Live button turns on real-time live events. If the end of the time-range includes the present, activating Live mode pulls in live events, so that you can see agents and tool calls in real-time.
- The dropdown selection below the Live button contains the following timeline view options: Past 15 minutes, Past 1 hour, Past 1 day, Past 1 week, Past two weeks, Past 1 month, or a Custom range.
- Selecting the or buttons allows you to scroll through the timeline history based for the time range you've selected. For example, if you select Past 1 week, you can go backward or forward in the history a week at a time.
Visualizations
The Agent graph shows agent-resource interactions in real-time. Agents are displayed as primary nodes, and tools are shown as secondary nodes. Active connections are shown as animated lines during tool execution.
Hover over or click on an agent in the Agent graph to display its activity in the Activity graph.
The Agent graph and Activity graph provide visual representations of agent resource connections and tool calls.
Usage and cost information
To help track resource consumption across all agents, the Token usage, the Estimated cost, the Total runtime (primary cost metric), and the Total runs are displayed beneath the Agent graph and Activity graph.
The usage, cost, and count metrics are displayed underneath the visualizations.
Agent tool usage, Agent total runtime, and Agent run count show global runtime statistics.
Individual agent view
To get to the individual agent view, click on the name of the agent in the agent sidebar. Each agent displays a Last status indicator (SUCCESS, MANUALLY_CANCELED, FAILURE, TIMED_OUT), a Last started timestamp, and a Last run duration for the last agent run that occurred.
Monitoring page for a single agent.
Each individual agent has the same monitoring dashboard as the all agents view, with the exception of the Agent total runtime and Agent run count global runtime statistics.
Agent run view
Individual agent runs are displayed when expanding the dropdown to the left of the agent name in the agent sidebar. Each individual agent run is represented by the date and time it occurred.
Monitoring page for a single agent run.
Clicking on the run displays the Run ID for that particular agent run, as well as the Status (SUCCESS, MANUALLY_CANCELED, FAILURE, TIMED_OUT), Started timestamp , and Run duration.
The usage and cost information, Token usage, Estimated cost, Total runtime, and Tool calls, is shown above the replay of the agent run, and the Current activity visual that contains the tool chips.