diff --git a/changes/16562-sql-deadlock b/changes/16562-sql-deadlock
new file mode 100644
index 0000000000..c4c725e435
--- /dev/null
+++ b/changes/16562-sql-deadlock
@@ -0,0 +1 @@
+Reduced the number of 'Deadlock found' errors seen by the server when multiple hosts share the same UUID
diff --git a/server/datastore/mysql/hosts.go b/server/datastore/mysql/hosts.go
index 54896dad12..f781dddc4d 100644
--- a/server/datastore/mysql/hosts.go
+++ b/server/datastore/mysql/hosts.go
@@ -4405,57 +4405,63 @@ func (ds *Datastore) UpdateHost(ctx context.Context, host *fleet.Host) error {
 			refetch_critical_queries_until = ?
 		WHERE id = ?
 	`
-	_, err := ds.writer(ctx).ExecContext(ctx, sqlStatement,
-		host.DetailUpdatedAt,
-		host.LabelUpdatedAt,
-		host.PolicyUpdatedAt,
-		host.NodeKey,
-		host.Hostname,
-		host.UUID,
-		host.Platform,
-		host.OsqueryVersion,
-		host.OSVersion,
-		host.Uptime,
-		host.Memory,
-		host.CPUType,
-		host.CPUSubtype,
-		host.CPUBrand,
-		host.CPUPhysicalCores,
-		host.HardwareVendor,
-		host.HardwareModel,
-		host.HardwareVersion,
-		host.HardwareSerial,
-		host.ComputerName,
-		host.Build,
-		host.PlatformLike,
-		host.CodeName,
-		host.CPULogicalCores,
-		host.DistributedInterval,
-		host.ConfigTLSRefresh,
-		host.LoggerTLSPeriod,
-		host.TeamID,
-		host.PrimaryIP,
-		host.PrimaryMac,
-		host.PublicIP,
-		host.RefetchRequested,
-		host.OrbitNodeKey,
-		host.RefetchCriticalQueriesUntil,
-		host.ID,
-	)
-	if err != nil {
-		return ctxerr.Wrapf(ctx, err, "save host with id %d", host.ID)
-	}
-	_, err = ds.writer(ctx).ExecContext(ctx, `
+	return ds.withRetryTxx(
+		ctx, func(tx sqlx.ExtContext) error {
+			_, err := tx.ExecContext(
+				ctx, sqlStatement,
+				host.DetailUpdatedAt,
+				host.LabelUpdatedAt,
+				host.PolicyUpdatedAt,
+				host.NodeKey,
+				host.Hostname,
+				host.UUID,
+				host.Platform,
+				host.OsqueryVersion,
+				host.OSVersion,
+				host.Uptime,
+				host.Memory,
+				host.CPUType,
+				host.CPUSubtype,
+				host.CPUBrand,
+				host.CPUPhysicalCores,
+				host.HardwareVendor,
+				host.HardwareModel,
+				host.HardwareVersion,
+				host.HardwareSerial,
+				host.ComputerName,
+				host.Build,
+				host.PlatformLike,
+				host.CodeName,
+				host.CPULogicalCores,
+				host.DistributedInterval,
+				host.ConfigTLSRefresh,
+				host.LoggerTLSPeriod,
+				host.TeamID,
+				host.PrimaryIP,
+				host.PrimaryMac,
+				host.PublicIP,
+				host.RefetchRequested,
+				host.OrbitNodeKey,
+				host.RefetchCriticalQueriesUntil,
+				host.ID,
+			)
+			if err != nil {
+				return ctxerr.Wrapf(ctx, err, "save host with id %d", host.ID)
+			}
+			_, err = tx.ExecContext(
+				ctx, `
 			UPDATE host_display_names
 			SET display_name=?
 			WHERE host_id=?`,
-		host.DisplayName(),
-		host.ID,
+				host.DisplayName(),
+				host.ID,
+			)
+			if err != nil {
+				return ctxerr.Wrapf(ctx, err, "update host_display_names for host id %d", host.ID)
+			}
+			return nil
+		},
 	)
-	if err != nil {
-		return ctxerr.Wrapf(ctx, err, "update host_display_names for host id %d", host.ID)
-	}
-	return nil
 }
 
 func (ds *Datastore) OSVersion(ctx context.Context, osVersionID uint, teamID *uint) (*fleet.OSVersion, *time.Time, error) {
diff --git a/server/datastore/mysql/mysql.go b/server/datastore/mysql/mysql.go
index 39faf5627e..685b9cc255 100644
--- a/server/datastore/mysql/mysql.go
+++ b/server/datastore/mysql/mysql.go
@@ -225,8 +225,13 @@ func withRetryTxx(ctx context.Context, db *sqlx.DB, fn txFn, logger log.Logger)
 		return nil
 	}
 
-	bo := backoff.NewExponentialBackOff()
-	bo.MaxElapsedTime = 5 * time.Second
+	expBo := backoff.NewExponentialBackOff()
+	// MySQL innodb_lock_wait_timeout default is 50 seconds, so transaction can be waiting for a lock for several seconds.
+	// Setting a higher MaxElapsedTime to increase probability that transaction will be retried.
+	// This will reduce the number of retryable 'Deadlock found' errors. However, with a loaded DB, we will still see
+	// 'Context cancelled' errors when the server drops long-lasting connections.
+	expBo.MaxElapsedTime = 1 * time.Minute
+	bo := backoff.WithMaxRetries(expBo, 5)
 	return backoff.Retry(operation, bo)
 }
 
@@ -333,8 +338,13 @@ func (ds *Datastore) writeChanLoop() {
 		case *fleet.Host:
 			item.errCh <- ds.UpdateHost(item.ctx, actualItem)
 		case hostXUpdatedAt:
-			query := fmt.Sprintf(`UPDATE hosts SET %s = ? WHERE id=?`, actualItem.what)
-			_, err := ds.writer(item.ctx).ExecContext(item.ctx, query, actualItem.updatedAt, actualItem.hostID)
+			err := ds.withRetryTxx(
+				item.ctx, func(tx sqlx.ExtContext) error {
+					query := fmt.Sprintf(`UPDATE hosts SET %s = ? WHERE id=?`, actualItem.what)
+					_, err := tx.ExecContext(item.ctx, query, actualItem.updatedAt, actualItem.hostID)
+					return err
+				},
+			)
 			item.errCh <- ctxerr.Wrap(item.ctx, err, "updating hosts label updated at")
 		}
 	}