Aggregates the input series under the ranges in the provided ranges series, and outputs a new series with one
point at the start of each range, where the value is the aggregation result. If a range has no points, no
point will be produced for it. If a range has no start, the point produced for it will be at the
start of the compute request range.
Converts a boolean series to a range series, where each contiguous run of true values becomes a
half-open range [start, end). The boolean series is treated as forward-filled: a value holds until the
next sample. For example, given timestamps [0, 1, 2, 3, 4] and values [true, true, true, false, false],
the result is a single range [0, 3). When openEnded is false and the series ends with true, the last
range is closed at the final timestamp, producing a zero-duration range (moment) for isolated true values.
The array is flattened out into a an arrow stream of bucketed primitive results, with an extra column to
indicate the index of the array that the bucket corresponds to.
The array is flattened out into a an arrow stream of bucketed primitive results, with an extra column to
indicate the index of the array that the bucket corresponds to.
Resamples the input series to a new resolution using interpolation.
Outputs data for timestamps corresponding to the defined frequency. Based on interpolation strategy,
determines range of timestamps to output data for and interpolates values where necessary.
Combines multiple enum series together and outputs a single series. The strategy to merge input values
with the same timestamp together is specified in the operation field.
Queries events and emits an enum series where each data point corresponds to an event and the enum value is
derived from a single event field (property, label, etc).
Keeps points from the base series where the boolean expression evaluates to true. Only outputs timestamps
where all input series have an entry for that timestamp, or a value can be filled using the interpolation
configuration.
Forward fill interpolation for resampling. Rounds the earliest and latest timestamp up to the nearest interval
multiple and generates timestamps at the given interval between the new earliest and latest timestamps.
For every timestamp in the resampled timestamps, takes the last known value before the timestamp as the value.
Contains a ComputeNodeResponse for each applicable grouping along with metadata describing the grouping.
All the contained ComputeNodeResponses are guaranteed to be of the same type.
The FE should try to pass in inputs in the order in which they should be
evaluated for optimization’s sake. Alternatively, we can let the user select
preconditions which they know to be cheaper to compute, which we will evaluate
first.
Filters points such that the log message in each point matches the given re2 regular expression.
Regular expression syntax: https://github.com/google/re2/wiki/Syntax.
Combines multiple log series together and outputs a single series. The strategy to merge input values with the
same timestamp together is specified in the operation field.
For every timestamp specified in the input series, outputs a value that is the maximum for that timestamp
across all input series.
Only outputs timestamps where all input series have an entry for that timestamp, or a value can be filled
using the interpolation configuration.
For every timestamp specified in the input series, outputs a value that is the mean for that timestamp
across all input series.
Only outputs timestamps where all input series have an entry for that timestamp, or a value can be filled
using the interpolation configuration.
For every timestamp specified in the input series, outputs a value that is the minimum for that timestamp
across all input series.
Only outputs timestamps where all input series have an entry for that timestamp, or a value can be filled
using the interpolation configuration.
Resamples the input series to a new resolution using interpolation.
Outputs data for timestamps corresponding to the defined frequency. Based on interpolation strategy,
determines range of timestamps to output data for and interpolates values where necessary.
Combines multiple numeric series together and outputs a single series. If the same timestamp is duplicated in
multiple input series, the output series will contain a single point with this timestamp. The strategy to
merge input values with the same timestamp together is specified in the operation field.
Expands ranges durations by adding temporal padding to the start, end, or both sides. Note that if the input
ranges contain a RangeValue, it will be dropped.
Specification of a page for a series. Returns raw undecimated points beginning nearest to the given page
token, advancing pageSize points in the time direction specified by the sign of the page size.
Produces a list of ranges for each point that is greater than its neighbors.
Peaks at edges are discarded, and continuous, multivalue, flat peaks will return all values.
Configures how long a condition has to be true for to output a time range, and what to use as the
start of the output range. For a point to be included in the output time range, both the minPoints and
minDuration conditions must be satisfied.
For every timestamp specified in the input series, outputs a value that is the product for that timestamp
across all input series.
Only outputs timestamps where all input series have an entry for that timestamp, or a value can be filled
using the interpolation configuration.
A scatter plot comprised of the values from one input series on the x-axis
and values from another input series on the y-axis, for all points within a requested time range.
For each timestamp, selects a single enum value from the 1D enum array at the specified index. If the index
is out of bounds for an array at a given timestamp, it is omitted.
For each timestamp, selects a single numeric value from the 1D numeric array at the specified index. If the
index is out of bounds for an array at a given timestamp, it is omitted.
Applies IIR-based signal filtering to input series. Includes low-pass, high-pass, band-pass, and band-stop
filters. Currently supports variable-order bidirectional Butterworth filters, with fixed-size padding based
on SciPy output.
Decimate by spatial trees.
Creates buckets by bisecting on each dimension, creating quadrants for 2d scatter and octants for 3d scatter.
Continues subdividing by prioritizing larger undivided buckets until reaching the max amount of buckets.
Outputs a set of ranges where the input series is stable. For each point, the min and max are calculated over
the specified lookback window, including the current point. A point is considered stable if its value does
not deviate from the calculated min and the max by more than the threshold and the total number of points
within the window is at least the specified amount. The threshold can be either fixed values or percentages
of the value. The lookback window must be strictly positive. The minimum points threshold defaults to 2.
Produces a list of ranges for which data does not exist for the specified duration or longer. Increases
window size by the specified staleness threshold on both ends to capture edge cases of data not currently
in view.
For every timestamp specified in the input series, outputs a value that is the sum for that timestamp
across all input series.
Only outputs timestamps where all input series have an entry for that timestamp, or a value can be filled
using the interpolation configuration.
A filter on a single tag key and set of values. For an IN operator, the tag key must match at least one of the
values in the set. For a NOT_IN operator, it must not match any of the values in the set.
Calculates the first discrete difference of the input series. For each point, computes the difference between
the current value and the previous value (current - previous). The first point in the input series is omitted
from the output since it has no previous point to compare against.
Computes the z-score for each series within a group-by family. For each group at each timestamp, the output is
(value - mean) / stddev, where mean and stddev are computed across all groups in the family at that timestamp,
using the interpolation configuration to align values across groups. When stddev is zero at a timestamp (e.g., all groups have the same value), that data point is excluded from the output.
If groupByTags is empty, all groups in the input series will be considered as a single family. This node only
produces meaningful results when the input has a group by.