diff --git a/src/main.rs b/src/main.rs
index 175a090..8eee7dc 100644
--- a/src/main.rs
+++ b/src/main.rs
@@ -5,6 +5,7 @@ use serde::{Deserialize, Serialize};
 use std::collections::HashMap;
 use std::convert::TryFrom;
 use std::time::{Duration, SystemTime};
+use std::fmt;
 
 const RESTIC_ALARM_BUCKET: &str = "restic-alarm-state";
 const RESTIC_ALARM_STATE_FILE: &str = "state.toml";
@@ -51,6 +52,30 @@ struct RepoConfig {
     alert_duration: u64,
 }
 
+struct AlertStatus {
+    alert: bool,
+    inactivity: Duration,
+    last_alert: Option<Duration>,
+}
+
+impl fmt::Display for AlertStatus {
+    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+        let days = |d: Duration| {
+            d.as_secs() / (3600 * 24)
+        };
+        let hours = |d: Duration| {
+            (d.as_secs() / 3600) % 24
+        };
+        write!(f, "{}: inactivity: {} days, {} hours, last_alert: {}",
+               (if self.alert { "ALERT" } else { "no alert" }),
+               days(self.inactivity), hours(self.inactivity),
+               match self.last_alert {
+                   None => "none".to_owned(),
+                   Some(d) => format!("{} days, {} hours", days(d), hours(d))
+               })
+    }
+}
+
 fn is_alert_needed(
     cfg: &RepoConfig,
     inactive_for: Duration,
@@ -200,17 +225,21 @@ async fn repo_last_snapshot(client: &s3::Client, repo: &str) -> eyre::Result<Opt
 // (e.g. if it fails to parse).
 // So the error must not be propagated to the toplevel, which would abort the
 // alert for remaining repositories; it should instead just be reported/logged.
-async fn check_repo(client: &s3::Client, state: &mut State, repo: &str) -> eyre::Result<()> {
+async fn check_repo(client: &s3::Client, state: &mut State, repo: &str) -> eyre::Result<Option<AlertStatus>> {
     let config = read_repo_config(client, repo).await?;
-    if let Some(last_snapshot) = repo_last_snapshot(client, repo).await? {
-        if is_alert_needed(&config, last_snapshot, state.last_alert(repo)) {
+    if let Some(inactivity) = repo_last_snapshot(client, repo).await? {
+        let last_alert = state.last_alert(repo);
+        let alert = is_alert_needed(&config, inactivity, last_alert);
+        if alert {
             println!("Sending alert to {} about bucket {}", config.email, repo);
-            send_email(&config.email, repo, last_snapshot).await?;
+            send_email(&config.email, repo, inactivity).await?;
         }
         state.update_last_alert(repo);
         write_state(client, state).await?;
+        Ok(Some(AlertStatus { alert, inactivity, last_alert }))
+    } else {
+        Ok(None)
     }
-    Ok(())
 }
 
 #[::tokio::main]
@@ -224,17 +253,20 @@ async fn main() -> eyre::Result<()> {
     let client = aws_sdk_s3::Client::from_conf(config);
 
     let repos = to_watch(&client).await?;
-    println!("Watching repos:");
+    println!("Watching {} repos", repos.len());
     for r in repos.iter() {
-        println!("{}", r);
+        println!("- {}", r);
     }
     let mut state = read_state(&client).await?;
     println!("Current state: {:?}", &state);
 
     for repo in repos {
         match check_repo(&client, &mut state, &repo).await {
-            Ok(()) => {
-                println!("{}: OK", &repo);
+            Ok(None) => {
+                println!("{}: no snapshot, skipping", &repo)
+            },
+            Ok(Some(status)) => {
+                println!("{}: {}", &repo, status);
             },
             Err(err) => {
                 // is this the best way to log the error?