Theme
Tool Analytics
When your bots call external APIs through tools, OmniBots records every execution. The Tool Analytics dashboard lets you monitor success rates, diagnose failures, and identify performance bottlenecks across all tools in your workspace.
Dashboard Overview
Navigate to Monitoring > Analytics > Tool Analytics to open the dashboard. The page displays a summary row of KPI cards followed by a per-tool performance table and trend charts.
KPI Cards
| Metric | Description |
|---|---|
| Total Executions | Number of tool calls made across all bots in the selected period |
| Success Rate | Percentage of executions that returned a successful response |
| Avg Response Time | Mean round-trip time from tool call initiation to response received |
| Error Count | Total number of failed executions in the selected period |
Tool analytics dashboard with KPI cards for total executions, success rate, average response time, and error count, plus a usage trend line chart
TIP
A sudden drop in success rate often indicates an upstream API issue. Click the error count card to jump directly to the error breakdown.
Per-Tool Performance Table
Below the KPI cards, a sortable table lists every tool registered in your workspace.
| Column | Description |
|---|---|
| Tool Name | The display name of the tool |
| Total Calls | Number of times this tool was invoked |
| Success Rate | Percentage of calls that succeeded |
| Avg Response Time | Mean response time in milliseconds |
| P95 Response Time | 95th percentile response time -- indicates worst-case latency |
| Errors | Count of failed calls |
| Last Called | Timestamp of the most recent invocation |
Click any row to open the detail view for that tool.
Per-tool performance table showing columns for tool name, total calls, success rate, average and P95 response time, error count, and last called timestamp, with error type breakdown panel below
Error Breakdown
The error breakdown panel categorizes failures by type so you can prioritize fixes.
| Error Type | Description |
|---|---|
| Timeout | The tool did not respond within the configured timeout window |
| Connection Error | OmniBots could not reach the tool's endpoint |
| HTTP 4xx | The tool returned a client error (e.g., 400 Bad Request, 401 Unauthorized, 404 Not Found) |
| HTTP 5xx | The tool returned a server error (e.g., 500 Internal Server Error, 503 Service Unavailable) |
| Schema Mismatch | The response did not match the expected output schema |
| Rate Limited | The tool's API returned a 429 Too Many Requests response |
Each error type shows the count and percentage of total errors. Click an error type to see the specific execution logs.
Usage Trends
A time-series chart shows tool execution volume over the selected date range. You can toggle between:
- All tools -- Aggregated view of total executions per hour or day.
- Single tool -- Select a specific tool from the dropdown to see its individual trend line.
The chart also supports overlaying the success rate as a secondary axis, so you can correlate volume spikes with failure rate changes.
Identifying Failing Tools
To quickly find tools that need attention, sort the performance table by Success Rate in ascending order. Tools with the lowest success rate appear at the top.
OmniBots also highlights tools that meet any of these conditions:
- Success rate below 90% in the last 24 hours
- Average response time above 5 seconds
- More than 10 consecutive failures
These tools are flagged with a warning indicator in the table.
WARNING
A tool with a consistently low success rate can degrade the user experience for every bot that uses it. Investigate and resolve tool failures promptly, or disable the tool temporarily while the upstream issue is addressed.
Retry Analytics
When a tool is configured with automatic retries, the dashboard tracks retry behavior.
| Metric | Description |
|---|---|
| Retry Rate | Percentage of calls that required at least one retry |
| Retries Before Success | Average number of retry attempts before a successful response |
| Retry Exhaustion Rate | Percentage of calls where all retry attempts failed |
High retry rates with eventual success may indicate intermittent upstream issues. High retry exhaustion rates signal a persistent problem.
Filtering and Export
Use the filters at the top of the page to narrow results by bot, tool, date range, or error type. The filtered dataset can be exported as CSV for further analysis.
Next Steps
- Voice Analytics -- Review voice-specific performance metrics
- Custom Reports -- Build reports that combine tool metrics with conversation data
- Reports Dashboard -- Access pre-built reports