Continuous Profiling Is Now Generally Available in Dash0
Blog post from Dash0
Dash0 has made continuous profiling generally available, integrating Polar Signals’ Parca-based technology with its OpenTelemetry-native observability platform to provide code-level CPU and performance visibility alongside logs, metrics, and traces. Using a low-overhead eBPF agent deployed per Kubernetes node, the feature samples executing code without application changes, restarts, or language-specific profilers, supporting major languages and offering both on-CPU and optional off-CPU analysis for blocked threads and system activity. Users can ingest profiles through OTLP and pprof, analyze CPU and memory behavior with query-time symbolization, flame graphs, caller-callee “sandwich” views, workload filtering, and derived function-level metrics, while Agent0 and connected coding agents can use production evidence to identify optimization opportunities. Profiling is disabled by default in the Dash0 Kubernetes Operator and requires elevated host access when enabled, while pricing is consumption-based at $0.60 per million samples with filtering controls through SignalControl. Planned developments include NVIDIA CUDA GPU profiling, stronger trace-to-profile correlation, and automated optimization recommendations through Agent0.
| Trend | Post Mentions | Total Month Mentions | Posts | Companies | MoM |
|---|---|---|---|---|---|
| OpenTelemetry | 7 | 468 | 108 | 50 | -46% |
| Kubernetes | 5 | 3,442 | 313 | 98 | -33% |
| MCP | 2 | 7,799 | 747 | 188 | -35% |
| Observability | 1 | 1,750 | 404 | 143 | -43% |
Use this post, company, and trend context to find content marketing opportunities, perform competitive analysis, or address product feature gaps via the Plushcap MCP server or the Plushcap API.