instr(server): Count transactions by platform, source (#2230)

Collect counts of transactions per `(platform, source,
contains_slashes)`. The goal of this metric is to verify a suspicion
that most `unknown` transactions from javascript projects are URL
transactions.

This metric replaces the now unused `transaction_source` metric, which
is unused and produced some high-cardinality tags.

relay-server/src/actors/processor.rs

@@ -25,12 +25,12 @@ use relay_filter::FilterStatKey;
 use relay_general::pii::{PiiAttachmentsProcessor, PiiConfigError, PiiProcessor};
 use relay_general::processor::{process_value, ProcessingState};
 use relay_general::protocol::Context::Trace;
-use relay_general::protocol::Contexts;
 use relay_general::protocol::{
  self, Breadcrumb, ClientReport, Csp, Event, EventType, ExpectCt, ExpectStaple, Hpkp, IpAddr,
  LenientString, Metrics, RelayInfo, Replay, ReplayError, SecurityReportType, SessionAggregates,
  SessionAttributes, SessionStatus, SessionUpdate, Timestamp, TraceContext, UserReport, Values,
 };
+use relay_general::protocol::{Contexts, TransactionSource};
 use relay_general::store::{
  ClockDriftProcessor, LightNormalizationConfig, MeasurementsConfig, TransactionNameConfig,
 };
@@ -66,7 +66,7 @@ use crate::extractors::RequestMeta;
 use crate::metrics_extraction::sessions::extract_session_metrics;
 use crate::metrics_extraction::transactions::extract_transaction_metrics;
 use crate::metrics_extraction::transactions::types::ExtractMetricsError;
-use crate::statsd::{RelayCounters, RelayHistograms, RelayTimers};
+use crate::statsd::{PlatformTag, RelayCounters, RelayHistograms, RelayTimers};
 use crate::utils::{
  self, get_sampling_key, log_transaction_name_metrics, ChunkedFormDataAggregator, FormDataIter,
  ItemAction, ManagedEnvelope, SamplingResult,
@@ -1832,10 +1832,19 @@ impl EnvelopeProcessorService {
  let source = event.get_transaction_source();
 
  metric!(
- counter(RelayCounters::EventTransactionSource) += 1,
- source = &source.to_string(),
- sdk = envelope.meta().client_name().unwrap_or("proprietary"),
- platform = event.platform.as_str().unwrap_or("other"),
+ counter(RelayCounters::EventTransaction) += 1,
+ source = match &source {
+ TransactionSource::Other(_) => "other",
+ source => source.as_str(),
+ },
+ platform =
+ PlatformTag::from(event.platform.as_str().unwrap_or("other")).as_str(),
+ contains_slashes =
+ if event.transaction.as_str().unwrap_or_default().contains('/') {
+ "true"
+ } else {
+ "false"
+ }
  );
 
  let span_count = event.spans.value().map(Vec::len).unwrap_or(0) as u64;

relay-server/src/statsd.rs

@@ -491,12 +491,12 @@ pub enum RelayCounters {
  ///
  /// This metric is tagged with:
  /// - `platform`: The event's platform, such as `"javascript"`.
- /// - `sdk`: The name of the Sentry SDK sending the transaction. This tag is only set for
- /// Sentry's SDKs and defaults to "proprietary".
  /// - `source`: The source of the transaction name on the client. See the [transaction source
  /// documentation](https://develop.sentry.dev/sdk/event-payloads/properties/transaction_info/)
  /// for all valid values.
- EventTransactionSource,
+ /// - `contains_slashes`: Whether the transaction name contains `/`. We use this as a heuristic
+ /// to represent URL transactions.
+ EventTransaction,
  /// The number of transaction events processed grouped by transaction name modifications.
  /// This metric is tagged with:
  /// - `source_in`: The source of the transaction name before normalization.
@@ -567,7 +567,7 @@ impl CounterMetric for RelayCounters {
  #[cfg(feature = "processing")]
  RelayCounters::ProcessingMessageProduced => "processing.event.produced",
  RelayCounters::EventProtocol => "event.protocol",
- RelayCounters::EventTransactionSource => "event.transaction_source",
+ RelayCounters::EventTransaction => "event.transaction",
  RelayCounters::TransactionNameChanges => "event.transaction_name_changes",
  RelayCounters::Requests => "requests",
  RelayCounters::ResponsesStatusCodes => "responses.status_codes",
@@ -578,3 +578,69 @@ impl CounterMetric for RelayCounters {
  }
  }
 }
+
+/// Low-cardinality platform that can be used as a statsd tag.
+pub enum PlatformTag {
+ Cocoa,
+ Csharp,
+ Edge,
+ Go,
+ Java,
+ Javascript,
+ Julia,
+ Native,
+ Node,
+ Objc,
+ Other,
+ Perl,
+ Php,
+ Python,
+ Ruby,
+ Swift,
+}
+
+impl PlatformTag {
+ pub fn as_str(&self) -> &str {
+ match self {
+ Self::Cocoa => "cocoa",
+ Self::Csharp => "csharp",
+ Self::Edge => "edge",
+ Self::Go => "go",
+ Self::Java => "java",
+ Self::Javascript => "javascript",
+ Self::Julia => "julia",
+ Self::Native => "native",
+ Self::Node => "node",
+ Self::Objc => "objc",
+ Self::Other => "other",
+ Self::Perl => "perl",
+ Self::Php => "php",
+ Self::Python => "python",
+ Self::Ruby => "ruby",
+ Self::Swift => "swift",
+ }
+ }
+}
+
+impl<S: AsRef<str>> From<S> for PlatformTag {
+ fn from(value: S) -> Self {
+ match value.as_ref() {
+ "cocoa" => Self::Cocoa,
+ "csharp" => Self::Csharp,
+ "edge" => Self::Edge,
+ "go" => Self::Go,
+ "java" => Self::Java,
+ "javascript" => Self::Javascript,
+ "julia" => Self::Julia,
+ "native" => Self::Native,
+ "node" => Self::Node,
+ "objc" => Self::Objc,
+ "perl" => Self::Perl,
+ "php" => Self::Php,
+ "python" => Self::Python,
+ "ruby" => Self::Ruby,
+ "swift" => Self::Swift,
+ _ => Self::Other,
+ }
+ }
+}