feat(profiling): Support inbound filters for profiles

CHANGELOG.md

@@ -11,6 +11,7 @@
 
 - Retain empty string values in `span.data` and `event.contexts.trace.data`. ([#4174](https://github.com/getsentry/relay/pull/4174))
 - Allow `sample_rate` to be float type when deserializing `DynamicSamplingContext`. ([#4181](https://github.com/getsentry/relay/pull/4181))
+- Support inbound filters for profiles. ([#4176](https://github.com/getsentry/relay/pull/4176))
 
 **Internal:**
 

relay-profiling/Cargo.toml

@@ -18,14 +18,17 @@ chrono = { workspace = true }
 data-encoding = { workspace = true }
 itertools = { workspace = true }
 relay-base-schema = { workspace = true }
+relay-dynamic-config = { workspace = true }
 relay-event-schema = { workspace = true }
+relay-filter = { workspace = true }
 relay-log = { workspace = true }
 relay-metrics = { workspace = true }
 relay-protocol = { workspace = true }
 serde = { workspace = true }
 serde_json = { workspace = true }
 serde_path_to_error = { workspace = true }
 thiserror = { workspace = true }
+url = { workspace = true }
 
 [dev-dependencies]
 insta = { workspace = true }

relay-profiling/src/error.rs

@@ -1,3 +1,5 @@
+use relay_filter::FilterStatKey;
+
 use thiserror::Error;
 
 #[derive(Debug, Error)]
@@ -32,6 +34,8 @@ pub enum ProfileError {
     DurationIsTooLong,
     #[error("duration is zero")]
     DurationIsZero,
+    #[error("filtered profile")]
+    Filtered(FilterStatKey),
 }
 
 impl ProfileError {

relay-profiling/src/lib.rs

@@ -40,10 +40,15 @@
 //! Relay will forward those profiles encoded with `msgpack` after unpacking them if needed and push a message on Kafka.
 
 use std::error::Error;
+use std::net::IpAddr;
 use std::time::Duration;
+use url::Url;
 
 use relay_base_schema::project::ProjectId;
-use relay_event_schema::protocol::{Event, EventId};
+use relay_dynamic_config::GlobalConfig;
+use relay_event_schema::protocol::{Csp, Event, EventId, Exception, LogEntry, Values};
+use relay_filter::{Filterable, ProjectFiltersConfig};
+use relay_protocol::{Getter, Val};
 use serde::Deserialize;
 use serde_json::Deserializer;
 
@@ -75,10 +80,55 @@ struct MinimalProfile {
     #[serde(alias = "profile_id", alias = "chunk_id")]
     event_id: ProfileId,
     platform: String,
+    release: Option<String>,
     #[serde(default)]
     version: sample::Version,
 }
 
+impl Filterable for MinimalProfile {
+    fn csp(&self) -> Option<&Csp> {
+        None
+    }
+
+    fn exceptions(&self) -> Option<&Values<Exception>> {
+        None
+    }
+
+    fn ip_addr(&self) -> Option<&str> {
+        None
+    }
+
+    fn logentry(&self) -> Option<&LogEntry> {
+        None
+    }
+
+    fn release(&self) -> Option<&str> {
+        self.release.as_deref()
+    }
+
+    fn transaction(&self) -> Option<&str> {
+        None
+    }
+
+    fn url(&self) -> Option<Url> {
+        None
+    }
+
+    fn user_agent(&self) -> Option<&str> {
+        None
+    }
+}
+
+impl Getter for MinimalProfile {
+    fn get_value(&self, path: &str) -> Option<Val<'_>> {
+        match path.strip_prefix("event.")? {
+            "release" => self.release.as_deref().map(|release| release.into()),
+            "platform" => Some(self.platform.as_str().into()),
+            _ => None,
+        }
+    }
+}
+
 fn minimal_profile_from_json(
     payload: &[u8],
 ) -> Result<MinimalProfile, serde_path_to_error::Error<serde_json::Error>> {
@@ -139,7 +189,13 @@ pub fn parse_metadata(payload: &[u8], project_id: ProjectId) -> Result<ProfileId
     Ok(profile.event_id)
 }
 
-pub fn expand_profile(payload: &[u8], event: &Event) -> Result<(ProfileId, Vec<u8>), ProfileError> {
+pub fn expand_profile(
+    payload: &[u8],
+    event: &Event,
+    client_ip: Option<IpAddr>,
+    filter_settings: &ProjectFiltersConfig,
+    global_config: &GlobalConfig,
+) -> Result<(ProfileId, Vec<u8>), ProfileError> {
     let profile = match minimal_profile_from_json(payload) {
         Ok(profile) => profile,
         Err(err) => {
@@ -156,6 +212,16 @@ pub fn expand_profile(payload: &[u8], event: &Event) -> Result<(ProfileId, Vec<u
             return Err(ProfileError::InvalidJson(err));
         }
     };
+
+    if let Err(filter_stat_key) = relay_filter::should_filter(
+        &profile,
+        client_ip,
+        filter_settings,
+        global_config.filters(),
+    ) {
+        return Err(ProfileError::Filtered(filter_stat_key));
+    }
+
     let transaction_metadata = extract_transaction_metadata(event);
     let transaction_tags = extract_transaction_tags(event);
     let processed_payload = match (profile.platform.as_str(), profile.version) {
@@ -200,7 +266,12 @@ pub fn expand_profile(payload: &[u8], event: &Event) -> Result<(ProfileId, Vec<u
     }
 }
 
-pub fn expand_profile_chunk(payload: &[u8]) -> Result<Vec<u8>, ProfileError> {
+pub fn expand_profile_chunk(
+    payload: &[u8],
+    client_ip: Option<IpAddr>,
+    filter_settings: &ProjectFiltersConfig,
+    global_config: &GlobalConfig,
+) -> Result<Vec<u8>, ProfileError> {
     let profile = match minimal_profile_from_json(payload) {
         Ok(profile) => profile,
         Err(err) => {
@@ -212,6 +283,16 @@ pub fn expand_profile_chunk(payload: &[u8]) -> Result<Vec<u8>, ProfileError> {
             return Err(ProfileError::InvalidJson(err));
         }
     };
+
+    if let Err(filter_stat_key) = relay_filter::should_filter(
+        &profile,
+        client_ip,
+        filter_settings,
+        global_config.filters(),
+    ) {
+        return Err(ProfileError::Filtered(filter_stat_key));
+    }
+
     match (profile.platform.as_str(), profile.version) {
         ("android", _) => android::chunk::parse(payload),
         (_, sample::Version::V2) => {
@@ -246,18 +327,39 @@ mod tests {
     #[test]
     fn test_expand_profile_with_version() {
         let payload = include_bytes!("../tests/fixtures/sample/v1/valid.json");
-        assert!(expand_profile(payload, &Event::default()).is_ok());
+        assert!(expand_profile(
+            payload,
+            &Event::default(),
+            None,
+            &ProjectFiltersConfig::default(),
+            &GlobalConfig::default()
+        )
+        .is_ok());
     }
 
     #[test]
     fn test_expand_profile_with_version_and_segment_id() {
         let payload = include_bytes!("../tests/fixtures/sample/v1/segment_id.json");
-        assert!(expand_profile(payload, &Event::default()).is_ok());
+        assert!(expand_profile(
+            payload,
+            &Event::default(),
+            None,
+            &ProjectFiltersConfig::default(),
+            &GlobalConfig::default()
+        )
+        .is_ok());
     }
 
     #[test]
     fn test_expand_profile_without_version() {
         let payload = include_bytes!("../tests/fixtures/android/legacy/roundtrip.json");
-        assert!(expand_profile(payload, &Event::default()).is_ok());
+        assert!(expand_profile(
+            payload,
+            &Event::default(),
+            None,
+            &ProjectFiltersConfig::default(),
+            &GlobalConfig::default()
+        )
+        .is_ok());
     }
 }

relay-profiling/src/outcomes.rs

@@ -17,5 +17,6 @@ pub fn discard_reason(err: ProfileError) -> &'static str {
         ProfileError::TooManyProfiles => "profiling_too_many_profiles",
         ProfileError::DurationIsTooLong => "profiling_duration_is_too_long",
         ProfileError::DurationIsZero => "profiling_duration_is_zero",
+        ProfileError::Filtered(_) => "profiling_filtered",
     }
 }

relay-server/src/services/processor.rs

@@ -1745,7 +1745,7 @@ impl EnvelopeProcessorService {
             // Process profiles before dropping the transaction, if necessary.
             // Before metric extraction to make sure the profile count is reflected correctly.
             let profile_id = match keep_profiles {
-                true => profile::process(state),
+                true => profile::process(state, &global_config),
                 false => profile_id,
             };
             // Extract metrics here, we're about to drop the event/transaction.
@@ -1772,7 +1772,7 @@ impl EnvelopeProcessorService {
 
         if_processing!(self.inner.config, {
             // Process profiles before extracting metrics, to make sure they are removed if they are invalid.
-            let profile_id = profile::process(state);
+            let profile_id = profile::process(state, &global_config);
             profile::transfer_id(state, profile_id);
 
             // Always extract metrics in processing Relays for sampled items.
@@ -1812,7 +1812,11 @@ impl EnvelopeProcessorService {
     ) -> Result<(), ProcessingError> {
         profile_chunk::filter(state);
         if_processing!(self.inner.config, {
-            profile_chunk::process(state, &self.inner.config);
+            profile_chunk::process(
+                state,
+                &self.inner.global_config.current(),
+                &self.inner.config,
+            );
         });
         Ok(())
     }

relay-server/src/services/processor/profile.rs

@@ -1,10 +1,12 @@
 //! Profiles related processor code.
+use std::net::IpAddr;
 
-use relay_dynamic_config::Feature;
+use relay_dynamic_config::{Feature, GlobalConfig};
 
 use relay_base_schema::events::EventType;
 use relay_config::Config;
 use relay_event_schema::protocol::{Contexts, Event, ProfileContext};
+use relay_filter::ProjectFiltersConfig;
 use relay_profiling::{ProfileError, ProfileId};
 use relay_protocol::Annotated;
 
@@ -88,7 +90,13 @@ pub fn transfer_id(
 }
 
 /// Processes profiles and set the profile ID in the profile context on the transaction if successful.
-pub fn process(state: &mut ProcessEnvelopeState<TransactionGroup>) -> Option<ProfileId> {
+pub fn process(
+    state: &mut ProcessEnvelopeState<TransactionGroup>,
+    global_config: &GlobalConfig,
+) -> Option<ProfileId> {
+    let client_ip = state.managed_envelope.envelope().meta().client_addr();
+    let filter_settings = &state.project_info.config.filter_settings;
+
     let profiling_enabled = state.project_info.has_feature(Feature::Profiling);
     let mut profile_id = None;
 
@@ -104,7 +112,14 @@ pub fn process(state: &mut ProcessEnvelopeState<TransactionGroup>) -> Option<Pro
                 return ItemAction::DropSilently;
             };
 
-            match expand_profile(item, event, &state.config) {
+            match expand_profile(
+                item,
+                event,
+                &state.config,
+                client_ip,
+                filter_settings,
+                global_config,
+            ) {
                 Ok(id) => {
                     profile_id = Some(id);
                     ItemAction::Keep
@@ -119,8 +134,21 @@ pub fn process(state: &mut ProcessEnvelopeState<TransactionGroup>) -> Option<Pro
 }
 
 /// Transfers transaction metadata to profile and check its size.
-fn expand_profile(item: &mut Item, event: &Event, config: &Config) -> Result<ProfileId, Outcome> {
-    match relay_profiling::expand_profile(&item.payload(), event) {
+fn expand_profile(
+    item: &mut Item,
+    event: &Event,
+    config: &Config,
+    client_ip: Option<IpAddr>,
+    filter_settings: &ProjectFiltersConfig,
+    global_config: &GlobalConfig,
+) -> Result<ProfileId, Outcome> {
+    match relay_profiling::expand_profile(
+        &item.payload(),
+        event,
+        client_ip,
+        filter_settings,
+        global_config,
+    ) {
         Ok((id, payload)) => {
             if payload.len() <= config.max_profile_size() {
                 item.set_payload(ContentType::Json, payload);
@@ -131,6 +159,9 @@ fn expand_profile(item: &mut Item, event: &Event, config: &Config) -> Result<Pro
                 )))
             }
         }
+        Err(relay_profiling::ProfileError::Filtered(filter_stat_key)) => {
+            Err(Outcome::Filtered(filter_stat_key))
+        }
         Err(err) => Err(Outcome::Invalid(DiscardReason::Profiling(
             relay_profiling::discard_reason(err),
         ))),