Time series

Time series charts show the change in your data over time which can help identify infrastructure issues, spikes, or dips in the data. This can be a simple line chart, an area chart, or a bar chart. A time series chart might be used to show the change in the volume of log events, error rates, latency, or other time-sensitive data.

Prerequisites

• Create an Axiom account.
• Create a dataset in Axiom where you send your data.
• Send data to your Axiom dataset.
• Create an empty dashboard.

Create a dashboard element

1. Go to the Dashboards tab and open the dashboard to which you want to add the time series.
2. Click Edit dashboard.
3. Click Add element in the top right corner.
4. Click Timeseries from the list.
5. Choose one of the following:
   • Click Builder to create your chart using a visual query builder. For more information, see Create chart using visual query builder.
   • Click APL to create your chart using the Axiom Processing Language (APL). Create a chart in the same way you create a chart in the Editor of the Query tab.
6. Optional: Configure the dashboard element.
7. Click Save.

The new element appears in your dashboard. At the bottom, click Save to save your changes to the dashboard.

Thresholds

Add a threshold line to compare your data with a target value, such as a latency objective or an error limit. The chart displays the threshold as a dashed horizontal line. Line, area, bar, and percentile charts support threshold lines. Each chart displays one threshold line.

To add a threshold line:

1. In the chart editor, run the query. The threshold fields appear after the query returns results. Before that, the Thresholds section displays "Run the query to set a threshold."
2. In the settings panel on the right, expand Thresholds. This section is after Y-axis.
3. If the chart has more than one aggregation, select an aggregation in Aggregation. Each aggregation has its own threshold.
4. In Value, enter the threshold in the units of the chart's values. For more information, see Threshold units.
5. Optional: In Color, select a color for the line. The default color is red.
6. Optional: In Highlight, select Above or Below to shade the time intervals where the data crosses the threshold. The default is None.

To delete a threshold line, click Remove threshold.

Threshold units

The Value field uses the units of the chart's values:

• Time: Select a unit from the list next to the field: ns, µs, ms, s, min, h, or d.
• Data size: Select a unit from the list next to the field. The list contains binary multiples (KiB, MiB, GiB) or decimal multiples (kB, MB, GB), depending on the unit of the field.
• Percent: Enter the percentage. For example, enter 95 for 95%.
• Other units: The field displays the unit, such as a currency symbol, before or after the value.

If the values have no unit, enter a number.

How Axiom displays thresholds

• Line: Axiom displays the threshold as a dashed line in the selected color. Axiom extends the y-axis to show the line, even when your data doesn't reach the threshold.
• Highlight: Axiom shades each time interval where a series that the chart displays is above or below the threshold. Series from a time comparison don't count. In percentile charts, Axiom shades the time interval if any percentile crosses the threshold. Axiom doesn't shade the newest time interval while it's incomplete.
• Tooltip: The tooltip displays a Threshold row. If you select a highlight, the tooltip marks each series value that crosses the threshold.

In stacked charts (area and bar charts), Axiom compares the stack total with the threshold, not the individual series. The tooltip displays a Total row and marks the total if it crosses the threshold. The editor displays the hint "Stacked charts compare the stack total."

If the query uses automatic bin sizes and the aggregation is a count or a sum, the values depend on the bin size. The bin size changes with the time range. In this case, the editor displays the following warning: "The bin size changes with the time range, so a count or sum threshold only fits one bin size. Set a fixed bin size to compare against a constant value." To compare the data with a constant value, set a fixed bin size in the query.

Example with Builder

Example with APL

['sample-http-logs']
| summarize count() by bin_auto(_time)