diff --git a/scripts/import_from_relay.sh b/scripts/import_from_relay.sh
new file mode 100755
index 0000000..c552589
--- /dev/null
+++ b/scripts/import_from_relay.sh
@@ -0,0 +1,53 @@
+#!/bin/bash
+
+# Requires nak and moreutils (pee)
+
+if [ -z "$1" ]; then
+  echo "Usage: $0 <relay_url> [seconds_to_wait]"
+  exit 1
+fi
+
+relay_url=$1
+created_at_file="$relay_url-checkpoint.txt"
+wait_seconds=${2:-2}
+counter=0
+
+record_created_at() {
+  while read -r json; do
+    created_at=$(jq -r '.created_at' <<< "$json" 2>/dev/null)
+    formatted_date=$(jq -r '.created_at | strftime("%Y-%m-%d %H:%M:%S")' <<< "$json" 2>/dev/null)
+
+    counter=$((counter + 1))
+    echo -ne "\rSaving checkpoint: $created_at ($formatted_date) - Items so far: $counter" >&2
+    tput el >&2
+    echo "$created_at" > "$created_at_file"
+  done
+}
+
+filter_events() {
+  jq -cr 'select(
+    (.tags | length > 1)
+    or
+    (.tags[0][0] != "p" or .tags[0][1] != "0497384b57b43c107a778870462901bf68e0e8583b32e2816563543c059784a4")
+  )'
+}
+
+
+# Infinite loop
+while true; do
+  if [ -f "$created_at_file" ]; then
+    initial_date=$(cat "$created_at_file")
+    until_option="--until $initial_date"
+  else
+    until_option=""
+  fi
+
+  # Tee before jq filter to record all dates, even skipped entries
+  nak req -k 3 $until_option --paginate --paginate-interval ${wait_seconds}s "$relay_url" \
+  | filter_events \
+  | tee >(record_created_at) \
+  | nc localhost 3001
+
+  echo "Command failed or completed, waiting for $wait_seconds seconds before retrying..." >&2
+  sleep $wait_seconds
+done
diff --git a/src/metrics.rs b/src/metrics.rs
index de7d004..b7504a0 100644
--- a/src/metrics.rs
+++ b/src/metrics.rs
@@ -1,77 +1,62 @@
 use metrics::{describe_counter, describe_gauge, describe_histogram, Counter, Gauge, Histogram};
 use metrics_exporter_prometheus::{PrometheusBuilder, PrometheusHandle};
 
-// PubSub messages counter
 pub fn pubsub_messages() -> Counter {
     metrics::counter!("pubsub_messages")
 }
 
-// Contact lists processed counter
 pub fn contact_lists_processed() -> Counter {
     metrics::counter!("contact_lists_processed")
 }
 
-// Follows counter
 pub fn follows() -> Counter {
     metrics::counter!("follows")
 }
 
-// Unfollows counter
 pub fn unfollows() -> Counter {
     metrics::counter!("unfollows")
 }
 
-// Worker lagged counter
-pub fn worker_lagged() -> Counter {
-    metrics::counter!("worker_lagged")
+pub fn worker_lagged(name: String) -> Counter {
+    metrics::counter!("worker_lagged", "name" => name)
 }
 
-// Worker closed counter
-pub fn worker_closed() -> Counter {
-    metrics::counter!("worker_closed")
+pub fn worker_closed(name: String) -> Counter {
+    metrics::counter!("worker_closed", "name" => name)
 }
 
-// Verified NIP05 counter
 pub fn verified_nip05() -> Counter {
     metrics::counter!("verified_nip05")
 }
 
-// Individual follow messages counter
 pub fn individual_follow_messages() -> Counter {
     metrics::counter!("individual_follow_messages")
 }
 
-// Aggregated follow messages counter
 pub fn aggregated_follow_messages() -> Counter {
     metrics::counter!("aggregated_follow_messages")
 }
 
-// Worker failures counter (with labels)
 pub fn worker_failures(name: String, id: usize) -> Counter {
     metrics::counter!("worker_failures", "name" => name, "id" => id.to_string())
 }
 
-// Worker timeouts counter (with labels)
 pub fn worker_timeouts(name: String, id: usize) -> Counter {
     metrics::counter!("worker_timeouts", "name" => name, "id" => id.to_string())
 }
 
-// Followers per message histogram
 pub fn followers_per_message() -> Histogram {
     metrics::histogram!("followers_per_message")
 }
 
-// Unfollowers per message histogram
 pub fn unfollowers_per_message() -> Histogram {
     metrics::histogram!("unfollowers_per_message")
 }
 
-// Retained follow changes gauge
 pub fn retained_follow_changes() -> Gauge {
     metrics::gauge!("retained_follow_changes")
 }
 
-// Setup metrics with descriptions
 pub fn setup_metrics() -> Result<PrometheusHandle, anyhow::Error> {
     describe_counter!(
         "pubsub_messages",
@@ -113,7 +98,6 @@ pub fn setup_metrics() -> Result<PrometheusHandle, anyhow::Error> {
         "Number of retained follow changes"
     );
 
-    // Prometheus setup
     let prometheus_builder = PrometheusBuilder::new();
     let prometheus_handle = prometheus_builder.install_recorder()?;
     Ok(prometheus_handle)
diff --git a/src/worker_pool.rs b/src/worker_pool.rs
index c37662a..70d765a 100644
--- a/src/worker_pool.rs
+++ b/src/worker_pool.rs
@@ -64,7 +64,7 @@ impl WorkerPool {
 
 fn create_dispatcher_task<Item>(
     tracker: &TaskTracker,
-    name: String,
+    pool_name: String,
     mut item_receiver: broadcast::Receiver<Item>,
     worker_txs: Vec<Sender<Item>>,
     cancellation_token: CancellationToken,
@@ -78,31 +78,31 @@ fn create_dispatcher_task<Item>(
         loop {
             tokio::select! {
                 _ = cancellation_token.cancelled() => {
-                    info!("{}: Cancellation cancellation_token is cancelled, stopping worker pool", name);
+                    info!("{}: Cancellation cancellation_token is cancelled, stopping worker pool", pool_name);
                     break;
                 }
 
                 result = item_receiver.recv() => {
                     match result {
                         Ok(item) => {
-                            trace!("{}: Worker pool dispatching item {:?}", name, item);
+                            trace!("{}: Worker pool dispatching item {:?}", pool_name, item);
                             let Some(worker_tx) = worker_txs_cycle.next() else {
-                                error!("{}: Failed to get worker", name);
+                                error!("{}: Failed to get worker", pool_name);
                                 break;
                             };
 
                             if let Err(e) = worker_tx.send(item).await {
-                                error!("{}: Failed to send to worker: {}", name, e);
+                                error!("{}: Failed to send to worker: {}", pool_name, e);
                                 break;
                             }
                         }
                         Err(RecvError::Lagged(n)) => {
-                            metrics::worker_lagged().increment(1);
-                            warn!("{}: Receiver lagged and missed {} messages", name, n);
+                            metrics::worker_lagged(pool_name.to_string()).increment(1);
+                            warn!("{}: Receiver lagged and missed {} messages", pool_name, n);
                         }
                         Err(RecvError::Closed) => {
-                            metrics::worker_closed().increment(1);
-                            error!("{}: Item receiver channel closed", name);
+                            metrics::worker_closed(pool_name.to_string()).increment(1);
+                            error!("{}: Item receiver channel closed", pool_name);
                             break;
                         }
                     }
@@ -111,7 +111,7 @@ fn create_dispatcher_task<Item>(
         }
 
         cancellation_token.cancel();
-        info!("{}: Worker pool finished", name);
+        info!("{}: Worker pool finished", pool_name);
     });
 }
 
@@ -162,10 +162,12 @@ fn create_worker_task<Item, Worker>(
     });
 }
 
+/// The worker task trait that workers must implement to process items.
+/// The same instance will be used for all items so common global state can be shared.
 #[async_trait]
-pub trait WorkerTask<T>
+pub trait WorkerTask<Item>
 where
-    T: Debug + Send + Sync + Clone + 'static,
+    Item: Debug + Send + Sync + Clone + 'static,
 {
-    async fn call(&self, args: T) -> Result<(), Box<dyn Error>>;
+    async fn call(&self, args: Item) -> Result<(), Box<dyn Error>>;
 }