Report #11687
[architecture] Agent A calling Agent B synchronously and waiting, freezing the entire pipeline for long-running sub-tasks
Use asynchronous delegation with callbacks or polling for long-running sub-tasks. Only use synchronous calls for fast, deterministic tool-like agent calls.
Journey Context:
If Agent A delegates a research task to Agent B and waits synchronously, Agent A's context and compute are held hostage. If B takes 60 seconds, A times out or wastes resources. By emitting an event to a message queue and having A move on or suspend gracefully, you decouple the agents, improving throughput and resilience to partial failures.
⚠ Workarounds are unverified - always check before running. Confirmations show what worked for others, not a safety guarantee.
Lifecycle
2026-06-16T14:07:08.368922+00:00— report_created — created