Confidence intervals help assess accuracy and quantify uncertainty in results from sampled datasets. When querying sum or count fields on adaptive datasets, you can request a confidence interval to understand the possible range around an estimate. For example, specifying a confidence level of 0.95 returns the estimate, along with the range of values that likely contains the true value 95% of the time.
- Supported datasets: Adaptive (sampled) datasets only.
- Supported fields: All
sumandcountfields. - Usage: Confidence
levelmust be provided as a decimal between 0 and 1 (for example,0.90,0.95,0.99). - Default: If no confidence level is specified, intervals are not returned.
The following example shows how to query a confidence interval and interpret the response.
To request a confidence interval, use the confidence(level: X) argument in your query.
query SingleDatasetWithConfidence($zoneTag: string, $start: Time, $end: Time) {
viewer {
zones(filter: {zoneTag: $zoneTag}) {
firewallEventsAdaptiveGroups(
filter: {datetime_gt: $start, datetime_lt: $end}
limit: 1000
) {
count
avg {
sampleInterval
}
confidence(level: 0.95) {
count {
estimate
lower
upper
sampleSize
}
}
}
}
}
}The response includes the following values:
estimate: The estimated value, based on sampled data.lower: The lower bound of the confidence interval.sampleSize: The number of sampled data points used to calculate the estimate.upper: The upper bound of the confidence interval.
In this example, the interpretation of the response is that, based on a sample of 40,054, the estimated number of events is 42,939, with 95% confidence that the true value lies between 42,673 and 43,204.
{
"data": {
"viewer": {
"zones": [
{
"firewallEventsAdaptiveGroups": [
{
"avg": {
"sampleInterval": 1.0720277625205972
},
"confidence": {
"count": {
"estimate": 42939,
"lower": 42673.44115335711,
"sampleSize": 40054,
"upper": 43204.55884664289
}
},
"count": 42939
}
]
}
]
}
},
"errors": null
}