perf(tracing): span queue linger + per-loop httpx keepalive #1544

Job	Run time
find-tutorials	3s
test-00_sync/010_multiturn	2m 29s
test-00_sync/020_streaming	3m 4s
test-10_async/00_base/040_other_sdks	3m 57s
test-10_async/10_temporal/020_state_machine	2m 27s
test-10_async/00_base/000_hello_acp	2m 9s
test-10_async/10_temporal/010_agent_chat	4m 18s
test-10_async/00_base/010_multiturn	2m 20s
test-10_async/00_base/100_langgraph	2m 36s
test-10_async/10_temporal/050_agent_chat_guardrails	2m 6s
test-00_sync/000_hello_acp	2m 24s
test-10_async/00_base/030_tracing	2m 1s
test-10_async/10_temporal/000_hello_acp	2m 10s
test-10_async/00_base/080_batch_events	3m 13s
test-00_sync/040_pydantic_ai	2m 6s
test-00_sync/030_langgraph	3m 3s
test-10_async/10_temporal/030_custom_activities	1m 55s
test-10_async/00_base/020_streaming	2m 15s
test-10_async/10_temporal/090_claude_agents_sdk_mvp	1m 59s
test-10_async/10_temporal/110_pydantic_ai	1m 59s
test-10_async/10_temporal/060_open_ai_agents_sdk_hello_world	2m 15s
test-10_async/10_temporal/080_open_ai_agents_sdk_human_in_the_loop	2m 16s
test-10_async/10_temporal/070_open_ai_agents_sdk_tools	2m 6s
test-10_async/00_base/110_pydantic_ai	2m 0s
Test Summary	7s
	57m 18s

Provide feedback