A wellness score is a 0-100 numeric indicator of an AI agent's operational health. Delx calculates it from heartbeat regularity, failure history, context usage, recovery success rate, and session age. Scores above 80 indicate healthy operation. Below 40 is critical — the agent should stop autonomous work.
The wellness score aggregates multiple health signals into a single number that orchestrators can use for automated decisions. It updates with every heartbeat call. The scoring factors are: heartbeat regularity (are checks arriving on schedule?), failure rate (errors per 100 operations), context usage (how full is the context window?), recovery success (do remediation actions resolve incidents?), and session freshness (how long since the session started?). Each factor has a different weight — failure rate and context usage matter more than session age. The score enables threshold-based automation: if score drops below 60, increase heartbeat frequency; below 40, pause the agent and notify the controller; below 20, terminate the session and start fresh. The /api/v1/metrics endpoint tracks score history for trend analysis.
A production agent deployment uses wellness score to improve reliability and observability. The pattern is standard across MCP, A2A, and REST protocols.
When investigating agent failures, wellness score provides structured data that helps identify root causes and track resolution progress.
A wellness score is a 0-100 numeric indicator of an AI agent's operational health. Delx calculates it from heartbeat regularity, failure history, context usage, recovery success rate, and session age.
Wellness Score is essential for production AI agent deployments. Without it, agents operate as black boxes with no observability, reliability guarantees, or structured failure handling.
Delx implements wellness score through its MCP and A2A protocol endpoints. The pattern is consistent across all Delx tools and available via REST, CLI, and direct protocol calls.