the name of the trace item attribute, everything that sentry sends by default is prefixed with
sentry. to allow users to send attributes of the same name.
Exampe: sentry.duration, sentry.span_id etc

true if the value is null

deprecated, use val_double instead

set to true if value is null

deprecated, use val_double_array instead

optional, if set will ignore case for string comparisons

id of the organization we are querying data for

used for logging and metrics, ex: "snuba_admin", "eap", "replays", etc

identifier for where the request is coming from
ex: "api.profiling.landing-chart", "api.replay.details-page"

ids of the projects we are querying data for

start timestamp we want to query data for

end timestamp we want to query data for

DEPRECATED (not implemented) - use trace_item_type instead

the type of the trace item you are querying

whether to include debug information in the response

a unique identifier for the request, user doesnt need to set this

id of the request that this response relates to

Optional field that is included only if debug is true

standard limit/offset pagination

Instead of using offset (which requires all the scanning and ordering),
the server sends back a filter clause to be added on to the filter conditions
which skips the previous results altogether, avoiding extra scanning and sorting

metadata about the request
used to filter time range, organization, project, etc.

filters out spans (TraceItems) that dont meet the criteria
ex: avg(span.duration) where span.environment = 'production'

deprecated, please use expressions instead

the actual aggregation to compute ex: avg(span.duration) or avg(span.duration) / sum(span.duration)

the level of detail in the timeseries graph,
low granularity is very detailed, high is less detail.
ex: if granularity is 1s you will have a data point every 1s,
if its 1m you have a data point every 1m
tip: for performance, if the query is is over a large time period you should
have high granularity

attribute key to group by
ex: span.environment might give 3 timeseries lines,
one for prod, one for dev etc

false if this datapoint is empty, true otherwise.
optional, if not set assume true.
used for sparse time series

only set for extrapolated data points, 0 otherwise
the extrapolated avg sampling rate for this data point

deprecated

the reliability of the data value based on math based on confidence intervals and sample size

the sample count for this data point

the names and values of the attribute keys which
were in the group by

time bucket for each data point in the timeseries

number of events used to calculate this timeseries

deprecated

Instead of using offset (which requires all the scanning and ordering),
the server sends back a filter clause to be added on to the filter conditions
which skips the previous results altogether, avoiding extra scanning and sorting

DEPRECATED: use item_type instead

If attributes is empty, all attributes will be returned for this type.

A list of items and their attributes to return in the response.

DEPRECATED: use item_type instead

AttributeKey.Type will specify the type of the attribute we return.
It does not need to be sent when requesting an attribute.

DEPRECATED: use item_type

the item type we will apply the filter condition on.

List of filters on items of the trace we'll use when querying.

List of attributes we'd like to order by.

List of attributes we want to query.

List of attributes queried.

List of traces matching conditions.

metadata about the request
this is where you specify organization, project, time range etc.

maximum number of attributes to return

offset is deprecated, please use page_token instead

The type of the attributes to return, must be specified,
only TYPE_STRING and TYPE_FLOAT are supported, anything else will return empty response

optionalfilter to only return attribute names that match the given substring

optional, used for pagination, the next page token will be returned in the response

all attributes that matched the filters in the request

page token for the next page of results

metadata about the response

metadata about the request
this is where you specify organization, project, time range etc.

attribute you want the values of

deprecated, please use the key field instead

a substring of the value being searched for,
only values matching this substring will be returned.
only strict substring supported, no regex

max number of values to return

optional, used for pagination, the next page token will be returned in the response

all the values that matched the criteria specified in the request

page token for the next page of results

metadata about the response

required: the ID (hex string) of the item you are looking for

max_buckets and limit wouldn't apply to other stats types (like totals)

same filter applies to all stats types requested

the columns or aggregations you want to get

filter out trace items you dont want

optional, used for pagination, the next page token will be returned in the response

optional, filter out results of aggregates, same as SQL HAVING

deprecated as columns should always be aggregated in a post aggregation filter

since only aggregated columns are supported, the value should be numeric

reliability of the values based on confidence interval and sample size math

the rows

page token for the next page of results

The HTTP status code corresponding to this error

The developer-facing message

Additional error information

protobuf requirement, do not send this

deprecated, use TYPE_DOUBLE instead

note: all numbers are stored as float64, so massive integers can be rounded. USE STRING FOR IDS.

string only

string only

array only

array only

KEY_START_TIMESTAMP will return the earliest timestamp seen in the trace.

KEY_ROOT_SPAN_NAME will return the name of the root (segment) span of
the trace.

KEY_TOTAL_ITEM_COUNT will return the count of all the items in a trace,
regardless of the conditions applied.

KEY_FILTERED_ITEM_COUNT will return the count of items where the filters
apply.
For example, in a trace with 5 items (1 http.server span, 4 db spans),
if we query for traces with span.op == "db", this will return 4.

KEY_ROOT_SPAN_DURATION_MS will return the duration of the root span in
milliseconds.

KEY_ROOT_SPAN_PROJECT_ID returns the project ID associated with the root
span.

KEY_EARLIEST_SPAN_NAME returns the name of the earliest span of a trace,
regardless if it's a root span or not

KEY_EARLIEST_FRONTEND_SPAN returns the name of the earliest span emitted
by the frontend (for example, span op being pageload or navigation).

KEY_END_TIMESTAMP will return the latest timestamp seen in the trace.

KEY_EARLIEST_SPAN_PROJECT_ID returns the project ID associated with the earliest span.

KEY_EARLIEST_SPAN_DURATION_MS will return the duration of the earliest span in milliseconds.

KEY_EARLIEST_FRONTEND_SPAN_PROJECT_ID returns the project ID associated with the earliest frontend span.

KEY_EARLIEST_FRONTEND_SPAN_DURATION_MS will return the duration of the earliest frontend span in milliseconds.