diff --git a/Makefile b/Makefile
index fe8127ac..8436b0b1 100644
--- a/Makefile
+++ b/Makefile
@@ -27,6 +27,7 @@ SCYLLA_TEST_FILTER := $(subst ${SPACE},${EMPTY},ClusterTests.*\
 :UseKeyspaceCaseSensitiveTests.*\
 :MetricsTests.Integration_Cassandra_ErrorsRequestTimeouts\
 :MetricsTests.Integration_Cassandra_Requests\
+:MetricsTests.Integration_Cassandra_StatsShardConnections\
 :-PreparedTests.Integration_Cassandra_PreparedIDUnchangedDuringReprepare\
 :HeartbeatTests.Integration_Cassandra_HeartbeatFailed\
 :ControlConnectionTests.Integration_Cassandra_TopologyChange\
@@ -68,6 +69,7 @@ CASSANDRA_TEST_FILTER := $(subst ${SPACE},${EMPTY},ClusterTests.*\
 :UseKeyspaceCaseSensitiveTests.*\
 :MetricsTests.Integration_Cassandra_ErrorsRequestTimeouts\
 :MetricsTests.Integration_Cassandra_Requests\
+:MetricsTests.Integration_Cassandra_StatsShardConnections\
 :-PreparedTests.Integration_Cassandra_PreparedIDUnchangedDuringReprepare\
 :PreparedTests.Integration_Cassandra_FailFastWhenPreparedIDChangesDuringReprepare\
 :HeartbeatTests.Integration_Cassandra_HeartbeatFailed\
diff --git a/include/cassandra.h b/include/cassandra.h
index f5882125..5de60fa3 100644
--- a/include/cassandra.h
+++ b/include/cassandra.h
@@ -1780,21 +1780,46 @@ cass_cluster_set_queue_size_event(CassCluster* cluster,
                                   unsigned queue_size));
 
 /**
- * Sets the number of connections made to each server in each
- * IO thread.
+ * Sets the number of connections opened by the driver to each host.
  *
- * <b>Default:</b> 1
+ * Notice that this overrides the number of connections per shard
+ * set by `cass_cluster_set_core_connections_per_shard()`.
+ *
+ * <b>Default:</b> 1 per shard (i.e. `cass_cluster_set_core_connections_per_shard(cluster, 1)`)
  *
  * @public @memberof CassCluster
  *
  * @param[in] cluster
  * @param[in] num_connections
  * @return CASS_OK if successful, otherwise an error occurred.
+ *
+ * @see cass_cluster_set_core_connections_per_shard()
  */
 CASS_EXPORT CassError
 cass_cluster_set_core_connections_per_host(CassCluster* cluster,
                                            unsigned num_connections);
 
+/**
+ * Sets the number of connections opened by the driver to each shard.
+ *
+ * Cassandra nodes are treated as if they have one shard.
+ * 
+ * This will override the `cass_cluster_set_core_connections_per_host`, if set.
+ *
+ * <b>Default:</b> 1
+ *
+ * @public @memberof CassCluster
+ *
+ * @param[in] cluster
+ * @param[in] num_connections
+ * @return CASS_OK if successful, otherwise an error occurred.
+ *
+ * @see cass_cluster_set_core_connections_per_host()
+ */
+CASS_EXPORT CassError
+cass_cluster_set_core_connections_per_shard(CassCluster* cluster,
+                                           unsigned num_connections);
+
 /**
  * Sets the maximum number of connections made to each server in each
  * IO thread.
diff --git a/scylla-rust-wrapper/src/cluster.rs b/scylla-rust-wrapper/src/cluster.rs
index ddc42ca3..eeba1744 100644
--- a/scylla-rust-wrapper/src/cluster.rs
+++ b/scylla-rust-wrapper/src/cluster.rs
@@ -12,7 +12,7 @@ use openssl::ssl::SslContextBuilder;
 use openssl_sys::SSL_CTX_up_ref;
 use scylla::client::execution_profile::ExecutionProfileBuilder;
 use scylla::client::session_builder::SessionBuilder;
-use scylla::client::{SelfIdentity, WriteCoalescingDelay};
+use scylla::client::{PoolSize, SelfIdentity, WriteCoalescingDelay};
 use scylla::frame::Compression;
 use scylla::policies::load_balancing::{
     DefaultPolicyBuilder, LatencyAwarenessBuilder, LoadBalancingPolicy,
@@ -25,7 +25,7 @@ use std::collections::HashMap;
 use std::convert::TryInto;
 use std::future::Future;
 use std::net::IpAddr;
-use std::num::NonZero;
+use std::num::{NonZero, NonZeroUsize};
 use std::os::raw::{c_char, c_int, c_uint};
 use std::str::FromStr;
 use std::sync::Arc;
@@ -47,6 +47,8 @@ const DEFAULT_MAX_SCHEMA_WAIT_TIME: Duration = Duration::from_millis(10000);
 const DEFAULT_SCHEMA_AGREEMENT_INTERVAL: Duration = Duration::from_millis(200);
 // - setting TCP_NODELAY is true
 const DEFAULT_SET_TCP_NO_DELAY: bool = true;
+// - connection pool size is 1 per shard
+const DEFAULT_CONNECTION_POOL_SIZE: PoolSize = PoolSize::PerShard(NonZeroUsize::new(1).unwrap());
 // - enabling write coalescing
 const DEFAULT_ENABLE_WRITE_COALESCING: bool = true;
 // - write coalescing delay
@@ -234,6 +236,7 @@ pub unsafe extern "C" fn cass_cluster_new() -> CassOwnedExclusivePtr<CassCluster
             .schema_agreement_interval(DEFAULT_SCHEMA_AGREEMENT_INTERVAL)
             .tcp_nodelay(DEFAULT_SET_TCP_NO_DELAY)
             .connection_timeout(DEFAULT_CONNECT_TIMEOUT)
+            .pool_size(DEFAULT_CONNECTION_POOL_SIZE)
             .write_coalescing(DEFAULT_ENABLE_WRITE_COALESCING)
             .write_coalescing_delay(DEFAULT_WRITE_COALESCING_DELAY)
             .keepalive_interval(DEFAULT_KEEPALIVE_INTERVAL)
@@ -454,6 +457,59 @@ pub unsafe extern "C" fn cass_cluster_set_connect_timeout(
     cluster.session_builder.config.connect_timeout = Duration::from_millis(timeout_ms.into());
 }
 
+#[unsafe(no_mangle)]
+pub unsafe extern "C" fn cass_cluster_set_core_connections_per_host(
+    cluster_raw: CassBorrowedExclusivePtr<CassCluster, CMut>,
+    num_connections: c_uint,
+) -> CassError {
+    let Some(cluster) = BoxFFI::as_mut_ref(cluster_raw) else {
+        tracing::error!(
+            "Provided null cluster pointer to cass_cluster_set_core_connections_per_host!"
+        );
+        return CassError::CASS_ERROR_LIB_BAD_PARAMS;
+    };
+
+    match NonZeroUsize::new(num_connections as usize) {
+        Some(non_zero_conns) => {
+            cluster.session_builder.config.connection_pool_size = PoolSize::PerHost(non_zero_conns);
+            CassError::CASS_OK
+        }
+        None => {
+            tracing::error!(
+                "Provided zero connections to cass_cluster_set_core_connections_per_host!"
+            );
+            CassError::CASS_ERROR_LIB_BAD_PARAMS
+        }
+    }
+}
+
+#[unsafe(no_mangle)]
+pub unsafe extern "C" fn cass_cluster_set_core_connections_per_shard(
+    cluster_raw: CassBorrowedExclusivePtr<CassCluster, CMut>,
+    num_connections: c_uint,
+) -> CassError {
+    let Some(cluster) = BoxFFI::as_mut_ref(cluster_raw) else {
+        tracing::error!(
+            "Provided null cluster pointer to cass_cluster_set_core_connections_per_shard!"
+        );
+        return CassError::CASS_ERROR_LIB_BAD_PARAMS;
+    };
+
+    match NonZeroUsize::new(num_connections as usize) {
+        Some(non_zero_conns) => {
+            cluster.session_builder.config.connection_pool_size =
+                PoolSize::PerShard(non_zero_conns);
+            CassError::CASS_OK
+        }
+        None => {
+            tracing::error!(
+                "Provided zero connections to cass_cluster_set_core_connections_per_shard!"
+            );
+            CassError::CASS_ERROR_LIB_BAD_PARAMS
+        }
+    }
+}
+
 #[unsafe(no_mangle)]
 pub unsafe extern "C" fn cass_cluster_set_coalesce_delay(
     cluster_raw: CassBorrowedExclusivePtr<CassCluster, CMut>,
diff --git a/src/testing_unimplemented.cpp b/src/testing_unimplemented.cpp
index 7be97ead..3843b8ac 100644
--- a/src/testing_unimplemented.cpp
+++ b/src/testing_unimplemented.cpp
@@ -65,10 +65,6 @@ CASS_EXPORT CassError cass_cluster_set_cloud_secure_connection_bundle_no_ssl_lib
 CASS_EXPORT void cass_cluster_set_constant_reconnect(CassCluster* cluster, cass_uint64_t delay_ms) {
   throw std::runtime_error("UNIMPLEMENTED cass_cluster_set_constant_reconnect\n");
 }
-CASS_EXPORT CassError cass_cluster_set_core_connections_per_host(CassCluster* cluster,
-                                                                 unsigned num_connections) {
-  throw std::runtime_error("UNIMPLEMENTED cass_cluster_set_core_connections_per_host\n");
-}
 CASS_EXPORT CassError cass_cluster_set_host_listener_callback(CassCluster* cluster,
                                                               CassHostListenerCallback callback,
                                                               void* data) {
diff --git a/tests/src/integration/objects/cluster.hpp b/tests/src/integration/objects/cluster.hpp
index a03f1af9..7b8f32cc 100644
--- a/tests/src/integration/objects/cluster.hpp
+++ b/tests/src/integration/objects/cluster.hpp
@@ -170,6 +170,19 @@ class Cluster : public Object<CassCluster, cass_cluster_free> {
     return *this;
   }
 
+    /**
+   * Assign the number of connections made to each shard
+   *
+   * NOTE: One extra connection is established (the control connection)
+   *
+   * @param connections Number of connection per shard (default: 1)
+   * @return Cluster object
+   */
+  Cluster& with_core_connections_per_shard(unsigned int connections = 1u) {
+    EXPECT_EQ(CASS_OK, cass_cluster_set_core_connections_per_shard(get(), connections));
+    return *this;
+  }
+
   /**
    * Sets credentials for plain text authentication
    *
diff --git a/tests/src/integration/tests/test_metrics.cpp b/tests/src/integration/tests/test_metrics.cpp
index be517261..006eb356 100644
--- a/tests/src/integration/tests/test_metrics.cpp
+++ b/tests/src/integration/tests/test_metrics.cpp
@@ -76,6 +76,31 @@ CASSANDRA_INTEGRATION_TEST_F(MetricsTests, ErrorsConnectionTimeouts) {
   EXPECT_GE(2u, metrics.errors.connection_timeouts);
 }
 
+/**
+ * This test ensures that the driver is reporting the number of connections
+ * when connection pool size is configured per shard.
+ */
+CASSANDRA_INTEGRATION_TEST_F(MetricsTests, StatsShardConnections) {
+  CHECK_FAILURE;
+
+  const unsigned int CONNS_PER_SHARD = 2;
+
+  Session session =
+      default_cluster().with_core_connections_per_shard(CONNS_PER_SHARD).connect();
+
+  size_t nr_hosts = explode(contact_points_, ',').size();
+  size_t nr_shards = Options::is_scylla() ? Options::smp() : 1;
+  size_t expected_connection_count = nr_hosts * nr_shards * CONNS_PER_SHARD;
+
+  CassMetrics metrics = session.metrics();
+  for (int i = 0; i < 100 && metrics.stats.total_connections < expected_connection_count; ++i) {
+    metrics = session.metrics();
+    msleep(100);
+  }
+
+  EXPECT_GE(metrics.stats.total_connections, expected_connection_count);
+}
+
 /**
  * This test ensures that the driver is reporting the proper timeouts for requests
  *