AI Lead Extraction is currently the main bottleneck, averaging 5.2 seconds per run. Raw Lead Capture is healthy at 273 milliseconds average. Overall the system needs attention: the failure rate is 16.7% and the retry rate is 25.0%.
Tasks today
0
Completed
6
Failed
2
Running
1
Clarification needed
1
Average response time
8.0 s
Median latency
8.4 s
P95 latency
10.8 s
Main bottleneck
AI Lead Extraction
5.2 s avg per run
Slowest workflow
AI Lead Extraction
5.2 s avg
Slowest recent task
TASK-1036
10.8 s
Total tokens
20.4k
Estimated cost
$0.01
Token-heavy agent
Lead Intake Router
18.1k tokens
Most-used tool
ollama_generate
20 calls
Most common error
crm no exact match
2 occurrences
Retry rate
25.0%
Failure rate
16.7%
Most-used model
llama3.1:8b
20 steps
Dashboard is running on demo data — Supabase is not configured yet. The view is healthy and the latest demo task was updated 42 days ago.
Tasks started per day
Average vs P95 response time (seconds)
Outcomes across all tasks
Total tokens consumed per day
Average seconds per run, by workflow
Errors and retries per day