Trace waterfalls for long-running agent work
2 Jul 2026 · 6 min read
Trace waterfalls reveal bottlenecks across model calls, tools, human waits, retries, and handoffs.
Long-running agent work often fails through delay rather than error. Waterfalls make waiting, retries, and parallelism visible.
AstraRunner treats this as an execution-platform concern: the run needs durable state, traceable decisions, cost visibility, and enough structure for teams to review what happened after the agent finishes.
Patterns that work
- Separate active model time from wait time.
- Show retries under the original step.
- Highlight approval delays and blocked states.
- Compare critical path across runs.
Implementation checklist
- Record start and end time per step.
- Distinguish queued, running, waiting, and blocked.
- Compute critical path duration.
- Expose slowest tools and roles.
Watch outs
- Wall-clock duration without state labels is misleading.
- Parallel fanout can hide total cost.
- Human wait time should not be blamed on model latency.
Related reading
How this fits the Astra stack
AstraRunner owns the operational path for agent work: orchestration, scheduling, approval gates, traces, costs, and role-agent handoffs. AstraMemory owns durable context and retrieval. AstraGenie turns those capabilities into product workflows for teams that want automation without rebuilding the platform layer.