Sep 14, 2026 - v1.103.0

New Features

• Datadog source: Bring telemetry in from Datadog with a new first-class source type.

• Dynatrace Enrichment processor: Enrich telemetry with Dynatrace context before it leaves the pipeline, with recommendations to help you place it correctly.

• Per-signal field processors: Copy Field, Move Field, and Concat have been rebuilt to work per signal type and to target timestamp, trace, and severity fields directly — not just attributes and body. The previous versions remain available and are marked deprecated.

• Resource Detection: per-detector attribute control: Choose exactly which resource attributes each detector contributes, instead of taking everything a detector produces.

• Lookup Fields: API source and persistent caching: The Lookup Fields processor can now pull its lookup data from an API endpoint, cache it to disk, and control reload interval and maximum cache size.

Improvements

• Overview page overhaul: A reworked overview with a new Fleet Health section, the top five configurations instead of the top three, standardized throughput and usage graphs, consistent health colors, and a cleaner, faster pipeline graph.

• "Collectors" everywhere: Agents are now consistently called collectors across the UI, audit trail, and URLs. Existing /agents links redirect automatically.

• Splunk HEC source: Added control over raw endpoint splitting strategy, receiver timeouts, maximum body size, and mTLS.

• REST API source: Added a configurable authorization header prefix.

• Performance: Faster configuration rendering and rollouts, better handling of large fleets under load, and a smoother pipeline graph on large configurations.

Bug Fixes

• Empty fleet selectors: A fleet with no selector criteria no longer matches every collector.

• Pipeline health metrics: Health metrics are now correct for collectors reporting on a non-default interval, and one-minute measurements no longer come back empty on the configuration page.