fix: Redact emails in Sentry error messages (#208)

reacherhq · Jul 8, 2021 · f73a209 · f73a209
1 parent a2690f1
commit f73a209
Show file tree

Hide file tree

Showing 6 changed files with 59 additions and 18 deletions.
diff --git a/Cargo.lock b/Cargo.lock
diff --git a/Cargo.toml b/Cargo.toml
@@ -10,7 +10,9 @@ async-recursion = "0.3"
 async-smtp = "0.4"
 check-if-email-exists = "0.8"
 env_logger = "0.8"
+lazy_static = "1.4"
 log = "0.4"
+regex = "1.5"
 sentry = "0.22"
 serde = { version = "1.0", features = ["derive"] }
 tokio = { version = "1.8", features = ["macros"] }

diff --git a/src/bin/heroku.rs b/src/bin/heroku.rs
@@ -14,16 +14,20 @@
 // You should have received a copy of the GNU Affero General Public License
 // along with this program.  If not, see <https://www.gnu.org/licenses/>.
 
-use reacher_backend::{routes::create_routes, sentry_util::setup_sentry};
+use reacher_backend::{
+	routes::create_routes,
+	sentry_util::{setup_sentry, CARGO_PKG_VERSION},
+};
 use std::{env, net::IpAddr};
 
 /// Run a HTTP server using warp.
 ///
 /// # Panics
 ///
-/// If at least one of the environment variables:
-/// - RCH_HTTP_HOST
-/// is malformed, then the program will panic.
+/// The program panics if at least one of the environment variables is
+/// malformed:
+/// - RCH_HTTP_HOST,
+/// - PORT.
 #[tokio::main]
 async fn main() -> Result<(), Box<dyn std::error::Error + Send + Sync>> {
 	env_logger::init();
@@ -43,6 +47,7 @@ async fn main() -> Result<(), Box<dyn std::error::Error + Send + Sync>> {
 		.unwrap_or(8080);
 	log::info!(target: "reacher", "Server is listening on {}:{}.", host, port);
 
+	log::info!(target: "reacher", "Running Reacher v{}", CARGO_PKG_VERSION);
 	warp::serve(routes).run((host, port)).await;
 	Ok(())
 }
diff --git a/src/routes/check_email/post.rs b/src/routes/check_email/post.rs
@@ -151,7 +151,8 @@ async fn check_email(body: EndpointRequest) -> Result<impl warp::Reply, warp::Re
 	};
 
 	// Log on Sentry the `is_reachable` field.
-	// FIXME We should definitely log this somehwere else than Sentry.
+	// We should definitely log this somewhere else than Sentry.
+	// TODO https://github.com/reacherhq/backend/issues/207
 	sentry_util::metrics(
 		format!("is_reachable={:?}", value.is_reachable),
 		retry_option,

diff --git a/src/sentry_util.rs b/src/sentry_util.rs
@@ -17,15 +17,22 @@
 //! Helper functions to send events to Sentry.
 
 use crate::routes::check_email::post::RetryOption;
+use lazy_static::lazy_static;
+use regex::Regex;
 use sentry::protocol::{Event, Level, Value};
 use std::{collections::BTreeMap, env};
 
 pub const CARGO_PKG_VERSION: &str = env!("CARGO_PKG_VERSION");
 
-/// Setup logging and Sentry.
-pub fn setup_sentry() -> sentry::ClientInitGuard {
-	log::info!(target: "reacher", "Running Reacher v{}", CARGO_PKG_VERSION);
+lazy_static! {
+	// Regex to extract emails from a string.
+	static ref RE: Regex =
+		Regex::new(r"[a-zA-Z0-9._-]+@(?P<domain>[a-zA-Z0-9._-]+\.[a-zA-Z0-9_-]+)")
+			.unwrap();
+}
 
+/// Setup Sentry.
+pub fn setup_sentry() -> sentry::ClientInitGuard {
 	// Use an empty string if we don't have any env variable for sentry. Sentry
 	// will just silently ignore.
 	let sentry = sentry::init(env::var("RCH_SENTRY_DSN").unwrap_or_else(|_| "".into()));
@@ -36,7 +43,7 @@ pub fn setup_sentry() -> sentry::ClientInitGuard {
 	sentry
 }
 
-/// If HEROKU_APP_NAME environment variable is set, add it to the sentry extra
+/// If HEROKU_APP_NAME environment variable is set, add it to the sentry `extra`
 /// properties.
 fn add_heroku_app_name(mut extra: BTreeMap<String, Value>) -> BTreeMap<String, Value> {
 	if let Ok(heroku_app_name) = env::var("HEROKU_APP_NAME") {
@@ -48,6 +55,7 @@ fn add_heroku_app_name(mut extra: BTreeMap<String, Value>) -> BTreeMap<String, V
 
 /// Helper function to send an Info event to Sentry. We use these events for
 /// analytics purposes (I know, Sentry shouldn't be used for that...).
+/// TODO https://github.com/reacherhq/backend/issues/207
 pub fn metrics(message: String, retry_option: RetryOption, duration: u128, domain: &str) {
 	log::info!("Sending info to Sentry: {}", message);
 
@@ -62,20 +70,19 @@ pub fn metrics(message: String, retry_option: RetryOption, duration: u128, domai
 		extra,
 		level: Level::Info,
 		message: Some(message),
-		// FIXME It seams that this doesn't work on Sentry, so I added it in
-		// the `extra` field above too.
 		release: Some(CARGO_PKG_VERSION.into()),
 		..Default::default()
 	});
 }
 
 /// Helper function to send an Error event to Sentry.
 pub fn error(message: String, result: Option<&str>, retry_option: Option<RetryOption>) {
-	log::debug!("Sending error to Sentry: {}", message);
-	let mut extra = BTreeMap::new();
+	let redacted_message = redact(message.as_str());
+	log::debug!("Sending error to Sentry: {}", redacted_message);
 
+	let mut extra = BTreeMap::new();
 	if let Some(result) = result {
-		extra.insert("CheckEmailOutput".into(), result.into());
+		extra.insert("CheckEmailOutput".into(), redact(result).into());
 	}
 	if let Some(retry_option) = retry_option {
 		extra.insert("retry_option".into(), retry_option.to_string().into());
@@ -85,10 +92,34 @@ pub fn error(message: String, result: Option<&str>, retry_option: Option<RetryOp
 	sentry::capture_event(Event {
 		extra,
 		level: Level::Error,
-		message: Some(message),
-		// FIXME It seams that this doesn't work on Sentry, so I added it in
-		// the `extra` field above too.
+		message: Some(redacted_message),
 		release: Some(CARGO_PKG_VERSION.into()),
 		..Default::default()
 	});
 }
+
+/// Function to parse emails inside a text, and replace them with
+/// `*@domain.com` for privacy reasons.
+fn redact(input: &str) -> String {
+	let result = RE.replace_all(input, "*@$domain");
+	result.into()
+}
+
+#[cfg(test)]
+mod tests {
+	use super::redact;
+
+	#[test]
+	fn test_redact() {
+		assert_eq!("*@gmail.com", redact("someone@gmail.com"));
+		assert_eq!(
+			"my email is *@gmail.com.",
+			redact("my email is someone@gmail.com.")
+		);
+		assert_eq!(
+			"my email is *@gmail.com., I repeat, my email is *@gmail.com.",
+			redact("my email is someone@gmail.com., I repeat, my email is someone@gmail.com.")
+		);
+		assert_eq!("someone @ gmail . com", redact("someone @ gmail . com"));
+	}
+}
diff --git a/tests/README.md b/tests/README.md
@@ -2,4 +2,4 @@
 
 Files in this folder are Reacher backend's integration tests. They need the following to work:
 
--   a working internet connection,
+-   a working internet connection.
Original file line number	Diff line number	Diff line change
Expand Up		@@ -2,4 +2,4 @@

		Files in this folder are Reacher backend's integration tests. They need the following to work:

		- a working internet connection,
		- a working internet connection.