relay_base_schema/data_category.rs
1//! Defines the [`DataCategory`] type that classifies data Relay can handle.
2
3use std::fmt;
4use std::str::FromStr;
5
6use serde::{Deserialize, Serialize};
7
8use crate::events::EventType;
9
10/// An error that occurs if a number cannot be converted into a [`DataCategory`].
11#[derive(Debug, PartialEq, thiserror::Error)]
12#[error("Unknown numeric data category {0} can not be converted into a DataCategory.")]
13pub struct UnknownDataCategory(pub u8);
14
15/// Classifies the type of data that is being ingested.
16#[derive(Clone, Copy, Debug, Eq, Hash, Ord, PartialEq, PartialOrd, Deserialize, Serialize)]
17#[serde(rename_all = "snake_case")]
18#[repr(i8)]
19pub enum DataCategory {
20 /// Reserved and unused.
21 Default = 0,
22 /// Error events and Events with an `event_type` not explicitly listed below.
23 Error = 1,
24 /// Transaction events.
25 Transaction = 2,
26 /// Events with an event type of `csp`, `hpkp`, `expectct` and `expectstaple`.
27 Security = 3,
28 /// An attachment. Quantity is the size of the attachment in bytes.
29 Attachment = 4,
30 /// Session updates. Quantity is the number of updates in the batch.
31 Session = 5,
32 /// Profile
33 ///
34 /// This is the category for processed profiles (all profiles, whether or not we store them).
35 Profile = 6,
36 /// Session Replays
37 Replay = 7,
38 /// DEPRECATED: A transaction for which metrics were extracted.
39 ///
40 /// This category is now obsolete because the `Transaction` variant will represent
41 /// processed transactions from now on.
42 TransactionProcessed = 8,
43 /// Indexed transaction events.
44 ///
45 /// This is the category for transaction payloads that were accepted and stored in full. In
46 /// contrast, `transaction` only guarantees that metrics have been accepted for the transaction.
47 TransactionIndexed = 9,
48 /// Monitor check-ins.
49 Monitor = 10,
50 /// Indexed Profile
51 ///
52 /// This is the category for indexed profiles that will be stored later.
53 ProfileIndexed = 11,
54 /// Span
55 ///
56 /// This is the category for spans from which we extracted metrics from.
57 Span = 12,
58 /// Monitor Seat
59 ///
60 /// Represents a monitor job that has scheduled monitor checkins. The seats are not ingested
61 /// but we define it here to prevent clashing values since this data category enumeration
62 /// is also used outside of Relay via the Python package.
63 MonitorSeat = 13,
64 /// User Feedback
65 ///
66 /// Represents a User Feedback processed.
67 /// Currently standardized on name UserReportV2 to avoid clashing with the old UserReport.
68 /// TODO(jferg): Rename this to UserFeedback once old UserReport is deprecated.
69 UserReportV2 = 14,
70 /// Metric buckets.
71 MetricBucket = 15,
72 /// SpanIndexed
73 ///
74 /// This is the category for spans we store in full.
75 SpanIndexed = 16,
76 /// ProfileDuration
77 ///
78 /// This data category is used to count the number of milliseconds per indexed profile chunk,
79 /// excluding UI profile chunks.
80 ProfileDuration = 17,
81 /// ProfileChunk
82 ///
83 /// This is a count of profile chunks received. It will not be used for billing but will be
84 /// useful for customers to track what's being dropped.
85 ProfileChunk = 18,
86 /// MetricSecond
87 ///
88 /// Reserved by billing to summarize the bucketed product of metric volume
89 /// and metric cardinality. Defined here so as not to clash with future
90 /// categories.
91 MetricSecond = 19,
92 /// Replay Video
93 ///
94 /// This is the data category for Session Replays produced via a video recording.
95 DoNotUseReplayVideo = 20,
96 /// This is the data category for Uptime monitors.
97 Uptime = 21,
98 /// Counts the number of individual attachments, as opposed to the number of bytes in an attachment.
99 AttachmentItem = 22,
100 /// LogItem
101 ///
102 /// This is the category for logs for which we store the count log events for users for measuring
103 /// missing breadcrumbs, and count of logs for rate limiting purposes.
104 LogItem = 23,
105 /// LogByte
106 ///
107 /// This is the category for logs for which we store log event total bytes for users.
108 LogByte = 24,
109 /// Profile duration of a UI profile.
110 ///
111 /// This data category is used to count the number of milliseconds per indexed UI profile
112 /// chunk.
113 ///
114 /// See also: [`Self::ProfileDuration`]
115 ProfileDurationUi = 25,
116 /// UI Profile Chunk.
117 ///
118 /// This data category is used to count the number of milliseconds per indexed UI profile
119 /// chunk.
120 ///
121 /// See also: [`Self::ProfileChunk`]
122 ProfileChunkUi = 26,
123 /// This is the data category to count Seer Autofix run events.
124 SeerAutofix = 27,
125 /// This is the data category to count Seer Scanner run events.
126 SeerScanner = 28,
127 /// PreventUser
128 ///
129 /// This is the data category to count the number of assigned Prevent Users.
130 PreventUser = 29,
131 /// PreventReview
132 ///
133 /// This is the data category to count the number of Prevent review events.
134 PreventReview = 30,
135 /// Size analysis
136 ///
137 /// This is the data category to count the number of size analyses performed.
138 /// 'Size analysis' a static binary analysis of a preprod build artifact
139 /// (e.g. the .apk of an Android app or MacOS .app).
140 /// When enabled there will typically be one such analysis per uploaded artifact.
141 SizeAnalysis = 31,
142 /// InstallableBuild
143 ///
144 /// This is the data category to count the number of installable builds.
145 /// It counts the number of artifacts uploaded *not* the number of times the
146 /// artifacts are downloaded for installation.
147 /// When enabled there will typically be one 'InstallableBuild' per uploaded artifact.
148 InstallableBuild = 32,
149 //
150 // IMPORTANT: After adding a new entry to DataCategory, go to the `relay-cabi` subfolder and run
151 // `make header` to regenerate the C-binding. This allows using the data category from Python.
152 // Rerun this step every time the **code name** of the variant is updated.
153 //
154 /// Any other data category not known by this Relay.
155 #[serde(other)]
156 Unknown = -1,
157}
158
159impl DataCategory {
160 /// Returns the data category corresponding to the given name.
161 pub fn from_name(string: &str) -> Self {
162 // TODO: This should probably use serde.
163 match string {
164 "default" => Self::Default,
165 "error" => Self::Error,
166 "transaction" => Self::Transaction,
167 "security" => Self::Security,
168 "attachment" => Self::Attachment,
169 "session" => Self::Session,
170 "profile" => Self::Profile,
171 "profile_indexed" => Self::ProfileIndexed,
172 "replay" => Self::Replay,
173 "transaction_processed" => Self::TransactionProcessed,
174 "transaction_indexed" => Self::TransactionIndexed,
175 "monitor" => Self::Monitor,
176 "span" => Self::Span,
177 "log_item" => Self::LogItem,
178 "log_byte" => Self::LogByte,
179 "monitor_seat" => Self::MonitorSeat,
180 "feedback" => Self::UserReportV2,
181 "user_report_v2" => Self::UserReportV2,
182 "metric_bucket" => Self::MetricBucket,
183 "span_indexed" => Self::SpanIndexed,
184 "profile_duration" => Self::ProfileDuration,
185 "profile_duration_ui" => Self::ProfileDurationUi,
186 "profile_chunk" => Self::ProfileChunk,
187 "profile_chunk_ui" => Self::ProfileChunkUi,
188 "metric_second" => Self::MetricSecond,
189 "replay_video" => Self::DoNotUseReplayVideo,
190 "uptime" => Self::Uptime,
191 "attachment_item" => Self::AttachmentItem,
192 "seer_autofix" => Self::SeerAutofix,
193 "seer_scanner" => Self::SeerScanner,
194 "prevent_user" => Self::PreventUser,
195 "prevent_review" => Self::PreventReview,
196 "size_analysis" => Self::SizeAnalysis,
197 "installable_build" => Self::InstallableBuild,
198 _ => Self::Unknown,
199 }
200 }
201
202 /// Returns the canonical name of this data category.
203 pub fn name(self) -> &'static str {
204 // TODO: This should probably use serde.
205 match self {
206 Self::Default => "default",
207 Self::Error => "error",
208 Self::Transaction => "transaction",
209 Self::Security => "security",
210 Self::Attachment => "attachment",
211 Self::Session => "session",
212 Self::Profile => "profile",
213 Self::ProfileIndexed => "profile_indexed",
214 Self::Replay => "replay",
215 Self::DoNotUseReplayVideo => "replay_video",
216 Self::TransactionProcessed => "transaction_processed",
217 Self::TransactionIndexed => "transaction_indexed",
218 Self::Monitor => "monitor",
219 Self::Span => "span",
220 Self::LogItem => "log_item",
221 Self::LogByte => "log_byte",
222 Self::MonitorSeat => "monitor_seat",
223 Self::UserReportV2 => "feedback",
224 Self::MetricBucket => "metric_bucket",
225 Self::SpanIndexed => "span_indexed",
226 Self::ProfileDuration => "profile_duration",
227 Self::ProfileDurationUi => "profile_duration_ui",
228 Self::ProfileChunk => "profile_chunk",
229 Self::ProfileChunkUi => "profile_chunk_ui",
230 Self::MetricSecond => "metric_second",
231 Self::Uptime => "uptime",
232 Self::AttachmentItem => "attachment_item",
233 Self::SeerAutofix => "seer_autofix",
234 Self::SeerScanner => "seer_scanner",
235 Self::PreventUser => "prevent_user",
236 Self::PreventReview => "prevent_review",
237 Self::SizeAnalysis => "size_analysis",
238 Self::InstallableBuild => "installable_build",
239 Self::Unknown => "unknown",
240 }
241 }
242
243 /// Returns true if the DataCategory refers to an error (i.e an error event).
244 pub fn is_error(self) -> bool {
245 matches!(self, Self::Error | Self::Default | Self::Security)
246 }
247
248 /// Returns the numeric value for this outcome.
249 pub fn value(self) -> Option<u8> {
250 // negative values (Internal and Unknown) cannot be sent as
251 // outcomes (internally so!)
252 (self as i8).try_into().ok()
253 }
254
255 /// Returns a dedicated category for indexing if this data can be converted to metrics.
256 ///
257 /// This returns `None` for most data categories.
258 pub fn index_category(self) -> Option<Self> {
259 match self {
260 Self::Transaction => Some(Self::TransactionIndexed),
261 Self::Span => Some(Self::SpanIndexed),
262 Self::Profile => Some(Self::ProfileIndexed),
263 _ => None,
264 }
265 }
266
267 /// Returns `true` if this data category is an indexed data category.
268 pub fn is_indexed(self) -> bool {
269 matches!(
270 self,
271 Self::TransactionIndexed | Self::SpanIndexed | Self::ProfileIndexed
272 )
273 }
274}
275
276impl fmt::Display for DataCategory {
277 fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
278 write!(f, "{}", self.name())
279 }
280}
281
282impl FromStr for DataCategory {
283 type Err = ();
284
285 fn from_str(string: &str) -> Result<Self, Self::Err> {
286 Ok(Self::from_name(string))
287 }
288}
289
290impl From<EventType> for DataCategory {
291 fn from(ty: EventType) -> Self {
292 match ty {
293 EventType::Default | EventType::Error | EventType::Nel => Self::Error,
294 EventType::Transaction => Self::Transaction,
295 EventType::Csp | EventType::Hpkp | EventType::ExpectCt | EventType::ExpectStaple => {
296 Self::Security
297 }
298 EventType::UserReportV2 => Self::UserReportV2,
299 }
300 }
301}
302
303impl TryFrom<u8> for DataCategory {
304 type Error = UnknownDataCategory;
305
306 fn try_from(value: u8) -> Result<Self, UnknownDataCategory> {
307 match value {
308 0 => Ok(Self::Default),
309 1 => Ok(Self::Error),
310 2 => Ok(Self::Transaction),
311 3 => Ok(Self::Security),
312 4 => Ok(Self::Attachment),
313 5 => Ok(Self::Session),
314 6 => Ok(Self::Profile),
315 7 => Ok(Self::Replay),
316 8 => Ok(Self::TransactionProcessed),
317 9 => Ok(Self::TransactionIndexed),
318 10 => Ok(Self::Monitor),
319 11 => Ok(Self::ProfileIndexed),
320 12 => Ok(Self::Span),
321 13 => Ok(Self::MonitorSeat),
322 14 => Ok(Self::UserReportV2),
323 15 => Ok(Self::MetricBucket),
324 16 => Ok(Self::SpanIndexed),
325 17 => Ok(Self::ProfileDuration),
326 18 => Ok(Self::ProfileChunk),
327 19 => Ok(Self::MetricSecond),
328 20 => Ok(Self::DoNotUseReplayVideo),
329 21 => Ok(Self::Uptime),
330 22 => Ok(Self::AttachmentItem),
331 23 => Ok(Self::LogItem),
332 24 => Ok(Self::LogByte),
333 25 => Ok(Self::ProfileDurationUi),
334 26 => Ok(Self::ProfileChunkUi),
335 27 => Ok(Self::SeerAutofix),
336 28 => Ok(Self::SeerScanner),
337 29 => Ok(Self::PreventUser),
338 30 => Ok(Self::PreventReview),
339 31 => Ok(Self::SizeAnalysis),
340 32 => Ok(Self::InstallableBuild),
341 other => Err(UnknownDataCategory(other)),
342 }
343 }
344}
345
346#[cfg(test)]
347mod tests {
348 use super::*;
349
350 #[test]
351 pub fn test_last_variant_conversion() {
352 // If this test fails, update the numeric bounds so that the first assertion
353 // maps to the last variant in the enum and the second assertion produces an error
354 // that the DataCategory does not exist.
355 assert_eq!(
356 DataCategory::try_from(32),
357 Ok(DataCategory::InstallableBuild)
358 );
359 assert_eq!(DataCategory::try_from(33), Err(UnknownDataCategory(33)));
360 }
361}