Timechart

The Timechart view renders results as a line chart with time on the x-axis. Best for visualizing trends, rates, and patterns over time.

Try it

datatable(timestamp:datetime, value:long, series:string) [
  datetime(2024-01-01 00:00), 10, "cpu",
  datetime(2024-01-01 01:00), 15, "cpu",
  datetime(2024-01-01 02:00), 12, "cpu",
  datetime(2024-01-01 03:00), 18, "cpu",
  datetime(2024-01-01 00:00), 50, "memory",
  datetime(2024-01-01 01:00), 55, "memory",
  datetime(2024-01-01 02:00), 48, "memory",
  datetime(2024-01-01 03:00), 60, "memory"
]
| render timechart

Column mapping

When no explicit xcolumn, ycolumns, or series properties are provided via render ... with (...), the timechart automatically detects column roles:

• X-axis — the first datetime column
• Y-axis — all int, long, or real columns (excluding the x-axis column)
• Series split — if there is exactly one remaining column (typically a string group-by column), each distinct value becomes a separate line

This means summarize count() by bin($time, 5m), service_name produces one line per service automatically — the datetime column is the x-axis, the count is the y-axis, and the service name splits the series.

For make-series results, the timechart detects array-valued columns (datetime arrays for the x-axis, numeric arrays for the y-axis) and renders each row as a separate series.

Use | render timechart with (xcolumn="...", series="...") to override the defaults. See Visualizations for all render properties.

Features

• Multiple series — each distinct value in the series column becomes a separate line
• Interactive legend — click legend entries to show/hide individual series
• Tooltips — hover over data points to see exact values
• Auto-render — use | render timechart to switch to this view automatically