Re:AI Agent Observability and Cost Attribution
by jasendo on Friday January 30, 2026 @03:24PM
(#46829396)
Cost attribution gets even worse when you factor in the bad decisions. Retries, abandoned tool calls, and cascading failures that still hit your bill but never show up in traces.
[ Parent
| on HN ]
Re:AI Agent Observability and Cost Attribution
by deborahjacob on Friday January 30, 2026 @04:21PM
(#46830063)
Iām building an OTel-based SDK that wraps the billable edges (entrypoint, LLM/tool clients, async publish/consume) and emits both traces for debugging and a lightweight event ledger for run/attempt lifecycle and call boundaries. I define the workflow + possible outcomes up front, then attribute all runs and attempts to the final outcome event to get the cost per outcome