feat(spans): Parse supabase span descriptions (#3153)

The description only contains the table name, e.g. `"from(users)"`.
In the future, we might want to parse `data.query` as well.

See
https://github.com/supabase-community/sentry-integration-js/blob/bae59de7e86c916d9396e2e2de55e6ac17e646d6/index.js#L259

Author: jjbayer

CHANGELOG.md

@@ -6,6 +6,7 @@
 
 - Extend GPU context with data for Unreal Engine crash reports. ([#3144](https://github.com/getsentry/relay/pull/3144))
 - Parametrize transaction in dynamic sampling context. ([#3141](https://github.com/getsentry/relay/pull/3141))
+- Parse & scrub span description for supabase. ([#3153](https://github.com/getsentry/relay/pull/3153))
 
 **Bug Fixes**:
 

relay-event-normalization/src/normalize/span/description/mod.rs

@@ -13,7 +13,8 @@ use relay_event_schema::protocol::Span;
 use url::Url;
 
 use crate::regexes::{
-    DB_SQL_TRANSACTION_CORE_DATA_REGEX, REDIS_COMMAND_REGEX, RESOURCE_NORMALIZER_REGEX,
+    DB_SQL_TRANSACTION_CORE_DATA_REGEX, DB_SUPABASE_REGEX, REDIS_COMMAND_REGEX,
+    RESOURCE_NORMALIZER_REGEX,
 };
 use crate::span::description::resource::COMMON_PATH_SEGMENTS;
 use crate::span::tag_extraction::HTTP_METHOD_EXTRACTOR_REGEX;
@@ -70,6 +71,11 @@ pub(crate) fn scrub_span_description(
                     // The description will only contain the entity queried and
                     // the query type ("User find" for example).
                     Some(description.to_owned())
+                } else if span_origin == Some("auto.db.supabase") {
+                    // The description only contains the table name, e.g. `"from(users)`.
+                    // In the future, we might want to parse `data.query` as well.
+                    // See https://github.com/supabase-community/sentry-integration-js/blob/master/index.js#L259
+                    scrub_supabase(description)
                 } else {
                     let (scrubbed, mode) = sql::scrub_queries(db_system, description);
                     if let sql::Mode::Parsed(ast) = mode {
@@ -141,6 +147,13 @@ fn scrub_core_data(string: &str) -> Option<String> {
     }
 }
 
+fn scrub_supabase(string: &str) -> Option<String> {
+    match DB_SUPABASE_REGEX.replace_all(string, "{%s}") {
+        Cow::Owned(scrubbed) => Some(scrubbed),
+        Cow::Borrowed(_) => None,
+    }
+}
+
 fn scrub_http(string: &str) -> Option<String> {
     let (method, url) = string.split_once(' ')?;
     if !HTTP_METHOD_EXTRACTOR_REGEX.is_match(method) {

relay-event-normalization/src/normalize/span/tag_extraction.rs

@@ -365,6 +365,12 @@ pub fn extract_tags(
             } else {
                 None
             }
+        } else if span.origin.as_str() == Some("auto.db.supabase") {
+            scrubbed_description.as_deref().map(|s| {
+                s.trim_start_matches("from(")
+                    .trim_end_matches(')')
+                    .to_owned()
+            })
         } else if span_op.starts_with("db") {
             span.description
                 .value()
@@ -1428,4 +1434,44 @@ LIMIT 1
             Some(&"Chrome".to_string())
         );
     }
+
+    #[test]
+    fn supabase() {
+        let json = r#"{
+            "description": "from(my_table00)",
+            "op": "db.select",
+            "origin": "auto.db.supabase",
+            "data": {
+                "query": [
+                    "select(*,other(*))",
+                    "in(something, (value1,value2))"
+                ]
+            }
+        }"#;
+
+        let span = Annotated::<Span>::from_json(json)
+            .unwrap()
+            .into_value()
+            .unwrap();
+
+        let tags = extract_tags(
+            &span,
+            &Config {
+                max_tag_value_size: 200,
+            },
+            None,
+            None,
+            false,
+            None,
+        );
+
+        assert_eq!(
+            tags.get(&SpanTagKey::Description).map(String::as_str),
+            Some("from(my_table{%s})")
+        );
+        assert_eq!(
+            tags.get(&SpanTagKey::Domain).map(String::as_str),
+            Some("my_table{%s}")
+        );
+    }
 }

relay-event-normalization/src/regexes.rs

@@ -81,3 +81,17 @@ pub static RESOURCE_NORMALIZER_REGEX: Lazy<Regex> = Lazy::new(|| {
 
 pub static DB_SQL_TRANSACTION_CORE_DATA_REGEX: Lazy<Regex> =
     Lazy::new(|| Regex::new(r"(?P<int>\d+)").unwrap());
+
+pub static DB_SUPABASE_REGEX: Lazy<Regex> = Lazy::new(|| {
+    Regex::new(
+        r"(?xi)
+        # UUIDs.
+        (?P<uuid>[a-f0-9]{8}-[a-f0-9]{4}-[a-f0-9]{4}-[a-f0-9]{4}-[a-f0-9]{12}) |
+        # Hexadecimal strings with more than 5 digits.
+        (?P<hex>[a-f0-9]{5}[a-f0-9]+) |
+        # Integer IDs with more than one digit.
+        (?P<int>\d\d+)
+        ",
+    )
+    .unwrap()
+});