{"slug": "a-timeout-is-not-proof-that-your-ai-agent-failed", "title": "A timeout is not proof that your AI agent failed", "summary": "Journey So Far published a developer practice scenario arguing that a timed-out AI agent request is not proof the underlying action failed, and that retrying blindly can duplicate orders, bookings or submissions. The scenario recommends separating conversation from action, using idempotency keys with an explicit contract, testing lost-response and concurrent-retry cases, and telling users honestly when completion cannot be confirmed. The site notes the scenario is fictional and project-authored, not a reported company interview.", "body_md": "An assistant submits a request, waits, and gets no reply. The tempting next step is “try again”. But a missing response does not tell us whether the underlying action happened. A second attempt could create a second order, booking or submission.\n\nThis makes a useful practice question for software developers moving into AI engineering: how would you find out what happened and decide what the assistant should do next?\n\nStart by separating the conversation from the action. The model proposing a retry is not evidence that repeating the operation is safe. Identify the operation, its request ID, whether it changes state, and whether the tool exposes a status query. An explicit permission check may also be needed before a consequential action.\n\nThen consider idempotency. A robust application can recognise that two requests represent the same intended action, rather than treating each transport attempt as a new instruction. That behaviour needs an explicit contract. Reusing a key with a different payload should not silently create ambiguity.\n\nTesting the happy path is not enough. Include a response lost after successful processing, two concurrent retries, a genuinely failed operation and an unavailable status service. For each case, specify what can be confirmed, what remains unknown and when the system should stop or escalate.\n\nFinally, explain the outcome honestly to the user. “I cannot confirm completion yet” may be a better response than claiming failure or creating another side effect. Record a bounded retry policy and a trace that makes the decision reviewable.\n\nThis is one possible reference approach, not a single employer's answer key. Before reading it, try explaining your own design in two minutes. What would you query? What would you refuse to repeat? Which failure would your test catch?\n\nWe make original scenarios like this at Journey So Far. Choose a role, practise a question, reveal a reference approach and explore free learning materials. No signup is needed.\n\nTechnical background, not the source of a copied interview question: [https://aws.amazon.com/builders-library/making-retries-safe-with-idempotent-APIs/](https://aws.amazon.com/builders-library/making-retries-safe-with-idempotent-APIs/)\n\nDisclosure: this article is from Journey So Far, the site behind these exercises. The scenario is fictional and project-authored; it is not a reported company interview.", "url": "https://wpnews.pro/news/a-timeout-is-not-proof-that-your-ai-agent-failed", "canonical_source": "https://dev.to/journeysofar/a-timeout-is-not-proof-that-your-ai-agent-failed-2e56", "published_at": "2026-11-19 03:35:23+00:00", "updated_at": "2026-11-19 03:41:43.541803+00:00", "lang": "en", "topics": ["ai-agents", "ai-tools", "developer-tools"], "entities": ["Journey So Far", "Amazon Web Services"], "also_reported_by": [], "takeaway": "Treat a timeout as unknown rather than failed: key each state-changing tool call with an idempotency key, query operation status before retrying, and test lost-response, concurrent-retry and unavailable-status cases.", "recommendation": {"advice": "Before retrying a timed-out agent action, query the operation's status by request ID and reuse an idempotency key so a duplicate attempt cannot create a second side effect.", "playbook": "agent-architecture", "playbook_url": "https://wpnews.pro/playbooks/agent-architecture"}, "alternates": {"html": "https://wpnews.pro/news/a-timeout-is-not-proof-that-your-ai-agent-failed", "markdown": "https://wpnews.pro/news/a-timeout-is-not-proof-that-your-ai-agent-failed.md", "text": "https://wpnews.pro/news/a-timeout-is-not-proof-that-your-ai-agent-failed.txt", "jsonld": "https://wpnews.pro/news/a-timeout-is-not-proof-that-your-ai-agent-failed.jsonld"}}