# Better metrics in Discover Services and new log visualizations

> Updates across Discover Services, Logs Explorer, and Control Plane improve accuracy, clarity, and workflow efficiency.

Source: https://last9.io/changelog/better-metrics-in-discover-services-and-new-log-visualizations/

### Improvements

- **Discover Services**: Operation Breakdown now calculates average calls per transaction directly from metrics, providing more accurate performance measurements - with a new 'Total time spent per minute' column and removes the less useful throughput column
- **Discover Exceptions**: Exception Traces can now be filtered by deployment environment (prod, qa, staging) when environment data is available, making it easier to isolate issues by stage
- **Logs Explorer**:
  - Logs table column order now persists across sessions, maintaining your preferred arrangement
  - Logs volume chart displays as a stacked bar chart grouped by severity, making error distribution over time easier to visualize
  - Top services table in Overview shows an "empty" label (in lighter gray) for logs without service names, and clicking it generates the correct search query
- **Control Plane > Scheduled Search**:
  - Alerts now handle very long queries more robustly with improved inspect link generation
  - Scheduled Search alert annotations (`query` and `query_type`) are now read-only in the UI, preventing accidental edits that could break alert functionality
- **Control Plane > Remapping**: Extract rule form now defaults to "Lines that match" instead of "All Lines", with a warning message when "All Lines" is selected to help prevent potential ingest delays

### Fixes

- Discover Hosts charts were not displaying correct x-axis values and were missing default values when metrics were unavailable
- Discover Applications "group by" selector was getting clipped at the edges, hiding some grouping options in the Performance and Errors tabs
- Discover Hosts queries for CPU and memory usage in statefulsets were causing chart rendering errors when metrics were missing
- Performance History was showing all gRPC calls instead of displaying the correct filtered data
- Log extraction scope form was not correctly defaulting to "Lines that match" when the extraction method was changed