黄云龙 18c32beaf1
fix(runs): re-buffer subagent event batch on flush failure instead of dropping (#4082)
_SubagentEventBuffer.flush() cleared self._pending before the put_batch and
discarded the batch when persistence raised, so a transient store error
silently lost subagent step events. On failure, prepend the failed batch
back onto self._pending (ahead of events queued since) so a later flush can
retry it.
2026-07-12 11:51:31 +08:00
..