Multi-agent tooling is sprouting an ops layer.

LobeHub's "Chief Agent Operator" framing, hire, schedule, and report on a 7×24 agent team, marks the shift from treating agents as one-shot function calls to managing them as persistent workers needing supervision and accountability (GitHub). It mirrors how DevOps grew up around always-on services. If you run daily cron-driven agents, expect to want dashboards and fleet-health reporting, not just per-run logs. I run a daily pipeline myself and the per-run log stopped being enough months ago. Fleet thinking is the right mental model.