Recurring agent failures are visible one run at a time while their common step remains hidden.
Developers need a fast path to repeated failure patterns, while variable traces make a misleading pattern costly.
Who you’d be doing this for
“I spent an hour flipping between six replays because every failure looked almost the same.”
Dina Esfahani · Application Developer
Owns an LLM-powered workflow and investigates production agent failures through Agent Tracing.
What is at stake
Developers are manually comparing several failed replays to find the same tool-call pattern. You have to weigh a useful shared view against labels that overstate what the traces prove.
Why it isn’t already fixed
Every obvious fix costs something else. That’s the part you’d have to decide.
- fast comparison vs. trace-query cost
- useful patterns vs. false causality
- narrow cohort vs. framework coverage
- automation-assisted discovery vs. verified evidence
Why Sentry
Agent Tracing already holds the runs, tool calls, and conversations developers need to compare when an LLM-powered service fails repeatedly.
Written with these in mind
Not your kind of problem? 20 more at Sentry, or browse every organization.
This is the setup. The work is inside.
Running it puts you in the room: the full situation and its constraints, stakeholders who push back in their own words, and the decisions that are yours to make. What you produce becomes a Day One Plan — work you can show someone instead of describing.