Platform expansion and AI-powered investigation workflows

Apple platform expansion and SAST improvements

Datadog RUM now supports watchOS and visionOS applications, extending mobile monitoring to Apple's newest platforms. This lets teams track real-user performance on Apple Watch and Vision Pro, validating that these environments have moved beyond experimentation into production workloads. Meanwhile, Datadog also published guidance on reducing SAST false positives using agentic evaluation with Bits Memories, addressing the common pain point of noisy static analysis results that slow down security reviews.

AI across the investigation lifecycle

Several vendors advanced AI-assisted debugging this week. Sentry's Seer can now hand off root cause analysis directly to GitHub Copilot, which generates a fix and pull request linked back in Sentry — eliminating manual context transfer. Datadog launched an Agent Observability toolset for debugging AI applications from inside coding agents like Claude Code, Cursor, and Codex CLI, including session classification and trace root-cause analysis skills. The company also published a post on accelerating investigations with AI in Incident Response, alongside its acquisition of Adaptive ML to build specialized AI agents fine-tuned on observability data.

Dash0 made Agent0 automations generally available, enabling autonomous investigation and fixes triggered by Slack, GitHub events, or failed checks, with guardrails requiring approval for write actions. The company also added spam filter effectiveness metrics showing dropped data over time, and improved chart rendering and list scrolling performance for high-cardinality visualizations.

Major SDK updates: Node.js v6 and Python fixes

The standout SDK release of the week is Datadog dd-trace-js v6.0.0, which drops Node.js 18 and 20 support while raising minimum versions for Cypress (12.0.0), Jest (28.0.0), Mocha (8.0.0), and Playwright (1.38.0). Alongside this major version, dd-trace-js 5.111.0 added OTLP trace metrics support and AppSec in-application WAF for AWS Lambda, while 5.110.0 introduced opt-in OpenTelemetry semantic conventions via DD_TRACE_OTEL_SEMANTICS_ENABLED.

On the Python side, multiple patch releases (4.9.4, 4.8.11, and 4.10.6) fix a critical issue where periodic thread restarts could block forked children from resuming, plus a rare crash on CPython before 3.12. The 4.10.6 release also fixes async generator wrapping on Python 3.11–3.14 and corrects under-reported token counts for OpenAI-compatible providers.

Sentry JavaScript 10.63.0 adds url.full to browser resource spans and extendIntegration plus isTracingSuppressed context strategy methods to the core package. Sentry Python 2.64.0 now streams gen_AI spans as v2 envelope items by default, preventing data loss on large payloads, though self-hosted instances should opt out. The Dash0 Operator version 0.146.0 adds node label collection and renames Intelligent Edge to Signal Control.

Observability infrastructure insights

Datadog published a detailed look at how it measures data completeness at scale across billions of payloads per second, and a post on using AI to build internal software delivery tools that keeps development workflows efficient. The Agent 7.80.4 release adds additional traces during SSI installation on Linux hosts for better debuggability.