TraceLab: Characterizing Coding Agent Workloads for LLM Serving
This work provides the first real-world trace of coding-agent workloads for serving-system analysis, addressing a critical data gap for researchers and engineers optimizing LLM serving infrastructure.
The paper introduces TraceLab, a dataset of ~4,300 coding-agent sessions with ~350,000 LLM steps and ~430,000 tool calls, revealing workload patterns like long autonomous loops, long contexts with short outputs, and high prefix cache hit rates, pointing to optimization opportunities for LLM serving.
Coding agents are rapidly becoming a major application of agentic LLMs, but serving them efficiently remains challenging. Progress on this challenge requires understanding real workload patterns, yet the data needed for such analysis is largely absent. Existing public traces and benchmarks do not capture real, day-to-day coding-agent usage across multiple agents and model families for serving-system analysis. To help fill this gap, we collect and release a trace of roughly 4,300 coding-agent sessions, containing about 350,000 LLM steps and 430,000 tool calls from our own day-to-day use of Claude Code and Codex. Our analysis shows that coding-agent workloads feature long autonomous loops, long contexts with short outputs, diverse and heavily-tailed tool calls, and high but imperfect prefix cache hit rates. These findings point to concrete opportunities for optimizing serving, including lower-overhead tool calling, append-length-aware prefill, semantic-aware tool-latency prediction, and improved KV-cache management around human-paced gaps. We release the dataset, trace collection pipeline, and analysis code at https://github.com/uw-syfi/TraceLab.git; the project website is https://tracelab.cs.washington.edu.