Cloudflare adds on-demand CPU/memory profiling with flamegraphs for Workers
Cloudflare Blog · rss · 2026-10-09
Cloudflare announced on-demand CPU and memory profiling for Workers and Durable Objects, viewable as interactive flamegraphs from the Observability page or CLI.
- Pick a version and duration, capture in production (needs sufficient traffic); enable source maps for readable TypeScript names; combine flamegraph with table view sorted by Samples.
- Real fix #1: R2 Worker's genericR2JsonReplacer walked the JSON tree redundantly inside the stringify replacer — made 2.7x faster; caching a repeated heavy metrics call saved more CPU.
- Real fix #2: a Worker OOMing at P999 133MB turned out to be a half-disabled Prometheus path eating 66.7% of allocations; removal brought P999 to 118MB under the 128MB limit.
- Explains the engineering behind routing profiling requests to the right data center/isolate/Durable Object.
More from Infra
- IFM open-sources RL360: 8,192 concurrent sandboxes and ~1.3M rollouts for agentic RL — rupspace · 2026-10-23
- SpaceX seeks access to 80,000 US cell tower sites; Polymarket puts carrier odds at 29% — Polymarket · 2026-10-23
- Modular open-sources MAX framework to free AI compute from single-vendor lock-in — clattner_llvm · 2026-10-23
- Claude Models on Google Cloud (Anthropic managed) is now generally available — claudeai · 2026-10-23
- Scaling GLM-5.3-Flash training to 1M token context, 64x over Megatron baseline — AndrewLBeam · 2026-10-23
- Perplexity's Portable Computer lands on AMD Ryzen AI Max, running local 27B model — AravSrinivas · 2026-10-23