Last9
Book demo
Mar 31st, 2026

Dashboard tags, per-request RUM correlation, and log pattern custom fields

Feature Improvements Fixes

TL;DR

Tag and filter dashboards to cut through the sprawl, correlate RUM requests with the exact backend span, invite new users directly from Settings, and extend log pattern detection to any field or URL path.

Dashboard tags, per-request RUM correlation, and log pattern custom fields

Dashboard tags

Dashboards now support tags for organizing and filtering. Group related dashboards (e.g. by service, team, or environment), save your tagged view, and jump back to it instantly instead of scrolling through every dashboard in the list.

Improvements

  • AI Assistant
    • MCP tokens now live under the Query section in Control Plane since hosted MCP has no write tools and feels more like a query credential
    • MCP query range follows your org's retention settings instead of a fixed 14-day cap
  • Control Plane
    • Mobile app origins (ios://, android://, tvos://, watchos:// with reverse-DNS bundle IDs) are now accepted for client monitoring tokens
  • Discover
    • Standardized the number of labels shown in tooltips and their sort order across service catalog views
  • Logs
    • Log pattern detection supports custom fields and URL paths — group logs by route or any attribute beyond the default body
  • RUM
    • Network request spans now carry a per-request traceId instead of sharing the view root span, so Related Logs by Trace ID surface only the investigated request's logs, not every request on the page
  • Settings
    • Admins can invite new users by email directly from the Users settings page, without going through a support request

Fixes

  • Dashboards
    • Chart breakdown view crashed with an undefined get error on some operations
    • Table column visibility couldn't be modified in custom dashboards
    • Chart tooltips covered the whole chart
    • A small-value line could disappear when the y-axis scale was dominated by larger series
  • Discover
    • Process metrics panels showed "No data" for Node.js services even when metrics were being sent
    • JVM process metrics were missing because the UI expected process_jvm_* but the metrics were jvm_*
    • Dependency tab showed 0 for services with valid throughput
    • Outgoing calls page showed no data for some services despite PromQL returning results
  • Control Plane
    • Logs drop page value picker was hard to use when picking values for attributes like msg
  • AI Assistant
    • MCP trace ID lookups returned no results even with a 7-day lookback
    • MCP service summary returned inflated latency for longer lookbacks
    • MCP OAuth consent page was unreadable in light theme
    • MCP Tokens Read section returned a 403 for some users
  • Settings
    • Permanent fix for an admin user error that had only been temporarily patched