perf(tracing): span queue linger + per-loop httpx keepalive #1543

Job	Run time
find-tutorials	4s
test-00_sync/040_pydantic_ai	2m 10s
test-10_async/00_base/020_streaming	2m 14s
test-10_async/00_base/000_hello_acp	4m 8s
test-00_sync/020_streaming	2m 45s
test-10_async/00_base/110_pydantic_ai	2m 7s
test-10_async/10_temporal/010_agent_chat	3m 12s
test-00_sync/010_multiturn	2m 53s
test-10_async/00_base/080_batch_events	3m 16s
test-10_async/00_base/040_other_sdks	4m 0s
test-00_sync/030_langgraph	3m 8s
test-10_async/00_base/010_multiturn	2m 5s
test-10_async/10_temporal/030_custom_activities	2m 5s
test-10_async/10_temporal/000_hello_acp	2m 4s
test-10_async/10_temporal/060_open_ai_agents_sdk_hello_world	2m 26s
test-10_async/10_temporal/080_open_ai_agents_sdk_human_in_the_loop	2m 29s
test-10_async/10_temporal/070_open_ai_agents_sdk_tools	2m 10s
test-10_async/10_temporal/050_agent_chat_guardrails	2m 3s
test-10_async/10_temporal/020_state_machine	2m 19s
test-10_async/00_base/100_langgraph	2m 13s
test-10_async/10_temporal/110_pydantic_ai	2m 17s
test-10_async/10_temporal/090_claude_agents_sdk_mvp	2m 0s
test-00_sync/000_hello_acp	2m 1s
test-10_async/00_base/030_tracing	2m 4s
Test Summary	5s
	58m 18s

Provide feedback