Returns a single current or past reported data value from the time series described by the expression. The returned value is displayed continuously across the chart, so you can use it as a reference value for comparing against the results of other queries.
|timeWindow||The report time of the data value to be returned, expressed as:
|expression||Expression describing the time series to return a data value from.|
at() standard time function returns the data value that was reported at a particular time (or, if such value is not available, the last reported value within a 1-hour time window preceding that particular time) by the time series described by the expression. The returned value is displayed continuously across the chart, so you can use it as a reference value for comparing against the results of other queries.
at() returns a separate result for each time series described by the expression.
You designate the data value of interest by specifying the time it was reported, typically as an interval before the current time. For example, you can specify
2h to obtain the data value that was reported 2 hours ago. In fact, you’ll see the value from 2 hours ago, even if you’re looking at a chart of past values from weeks or months ago.
As an alternative, you can specify the time relative to your chart. For example, you can use
"start" to obtain the data value that was reported at the beginning of your current chart window. In fact, if you’re looking at a chart of past values,
at() to return the earliest of those past values.
When live data is reported,
at() adjusts the returned value as necessary to reflect any changes in the data reported by the time series. Adjustments can occur whenever the chart is updated (every 30 seconds). For example, consider a time series that reports a different value every 2 minutes. If you run a query with
at() at 8:00am to return the value reported 2 hours ago (at 6:00am), the result is automatically adjusted around 8:02am to show the value that was reported at 6:02am.
Here’s a query that shows the maximum request latency across all sources and point tags.
Now we’d like to see how these maximums compare to the maximum request latency that was reported 2 weeks ago. We add a second query that applies
at() to the original query. From this chart, we can see some increases in the maximum latency over 2 weeks ago.
at() to the results of
integral() may produce unexpected results because
integral() considers time window boundaries inclusive. Use the
msum() function with a
1vw time window argument instead.