Machine‑to‑machine traffic now dominates networks
New analysis finds machine‑to‑machine traffic accounts for roughly 70% of enterprise network activity, driven by IoT, microservices, and GenAI — pressuring legacy observability built for human sessions. The shift means agent platforms must capture and analyze high‑volume M2M events, API usage, and anomalous bot behavior. (webpronews.com)
Akamai’s 2024 web-scraping report measured bots at 42.1% of overall web traffic and found 65.3% of that bot traffic was malicious. (staging-www.akamai.com) Cloudflare’s leadership has warned that AI-driven crawler and agent activity is accelerating, with CEO Matthew Prince saying online bot traffic could exceed human traffic by 2027. (techcrunch.com) Cloudflare’s 2025 telemetry also attributed roughly 4.2% of web traffic to generative‑AI bots and noted search crawlers like Googlebot remain a large share of automated requests. (how2shout.com) Industry observability research shows platform teams are already feeling telemetry growth: New Relic’s 2024 Observability Forecast surveyed 1,700 technology professionals and flagged rising telemetry volumes and tool sprawl as top challenges. (newrelic.com) OpenTelemetry guidance recommends head/tail sampling once trace rates exceed ~1,000 traces per second and encourages pipeline-level sampling and rule-based retention to avoid unbounded data costs. (opentelemetry.io) Agent-specific monitoring frameworks call for capturing prompts, tool calls, and multi-step reasoning chains as discrete observability artifacts, with recent work (AgentOps) and LangChain production guides outlining traceability, fidelity checks, and anomaly signals for LLM agents. (arxiv.org) Security reports and network vendors stress the operational consequence: IBM X-Force documents increased automation in reconnaissance and API abuse, while vendors like Nokia and Cisco say AI workloads are changing traffic patterns and require updated connectivity and network planning. (ibm.com)
