The logs behind every aggregate query, drag-to-zoom alerts, and mobile RUM parity

Feature Improvements Fixes

TL;DR

See the raw logs behind any aggregate chart, drag-to-zoom on alert timelines, create metrics from traces, and mobile RUM custom events now reach Logs.

The logs behind every aggregate query, drag-to-zoom alerts, and mobile RUM parity

When a logs pipeline includes an aggregate step (like count() over 5m), the UI used to show only the count chart — to see the actual log lines behind a spike you had to manually strip the aggregate step back out. Now the top 5 contributing log lines appear right below the chart, most recent first, each with its timestamp, severity, and message, plus a “View all matching logs” link that opens the full filter-only query. This is especially handy when you land on the logs screen from an aggregate-based alert.

Improvements

  • Traces: Non-timeslice trace aggregations now have a Create Metric button, matching the logs-to-metrics flow
  • Alerting: The alert rule timeline supports drag-to-zoom, instead of only selecting a range from the date picker
  • Applications (RUM)
    • Mobile custom events (addEvent) now emit OTLP logs in addition to trace events, so mobile funnel markers show up in Logs the same way browser RUM does
    • Browser RUM can now disable traceparent headers when needed, with documentation
  • Discover Kubernetes: Deployment Events now include ConfigMap and HPA subjects, and detail pages (Deployment, Pod, Node, StatefulSet) show the cluster name in the breadcrumb

Fixes

  • Logs: The log chart tooltip overflowed its container
  • Alerting
    • Scheduled-search alerts saved with a custom aggregate alias (not result) could silently never fire
    • The scheduled-search Slack log preview queried a window offset roughly 60 seconds from the evaluation window, so it could miss matching logs from the last minute
  • Dashboards
    • Time series charts didn’t render in bar view on some tenants
    • PromQL empty-state operators were incorrect, and a Kubernetes deployment fetch was duplicated
  • Discover Services: The Apdex chart showed 0 instead of an empty state when there were no requests