#include <iostream>
#include <iomanip>
#include <cassert>
#include <vector>
#include <algorithm>
#include <utility>
#include <numeric>
#include <tuple>
#include <ranges>
namespace ranges = std::ranges;
namespace views = std::views;
// #include "Src/Number/IntegerDivision.hpp"
// #include "Src/Utility/BinarySearch.hpp"


#include <cstdint>
#include <cstddef>

namespace zawa {

using i16 = std::int16_t;
using i32 = std::int32_t;
using i64 = std::int64_t;
using i128 = __int128_t;

using u8 = std::uint8_t;
using u16 = std::uint16_t;
using u32 = std::uint32_t;
using u64 = std::uint64_t;

using usize = std::size_t;

} // namespace zawa

#include <iterator>
#include <limits>

namespace zawa {

template <class T>
class CompressedSequence {
public:

    static constexpr u32 NotFound = std::numeric_limits<u32>::max();

    CompressedSequence() = default;

    template <class InputIterator>
    CompressedSequence(InputIterator first, InputIterator last) : comped_(first, last), f_{} {
        std::sort(comped_.begin(), comped_.end());
        comped_.erase(std::unique(comped_.begin(), comped_.end()), comped_.end());
        comped_.shrink_to_fit();
        f_.reserve(std::distance(first, last));
        for (auto it{first} ; it != last ; it++) {
            f_.emplace_back(std::distance(comped_.begin(), std::lower_bound(comped_.begin(), comped_.end(), *it)));
        }
    }

    CompressedSequence(const std::vector<T>& A) : CompressedSequence(A.begin(), A.end()) {}

    inline usize size() const noexcept {
        return comped_.size();
    }

    u32 operator[](const T& v) const {
        return std::distance(comped_.begin(), std::lower_bound(comped_.begin(), comped_.end(), v));
    }

    u32 upper_bound(const T& v) const {
        return std::distance(comped_.begin(), std::upper_bound(comped_.begin(), comped_.end(), v));
    }

    u32 find(const T& v) const {
        u32 i = std::distance(comped_.begin(), std::lower_bound(comped_.begin(), comped_.end(), v));
        return i == comped_.size() or comped_[i] != v ? NotFound : i;
    }

    bool contains(const T& v) const {
        u32 i = std::distance(comped_.begin(), std::lower_bound(comped_.begin(), comped_.end(), v));
        return i < comped_.size() and comped_[i] == v;
    }

    u32 at(const T& v) const {
        u32 res = find(v);
        assert(res != NotFound);
        return res;
    }

    inline u32 map(u32 i) const noexcept {
        assert(i < f_.size());
        return f_[i];
    }

    inline T inverse(u32 i) const noexcept {
        assert(i < size());
        return comped_[i];
    }

    inline std::vector<T> comped() const noexcept {
        return comped_;
    }

private:

    std::vector<T> comped_;

    std::vector<u32> f_;

};

} // namespace zawa
// #include "Src/Sequence/RunLengthEncoding.hpp"
// #include "Src/Algebra/Group/AdditiveGroup.hpp"
// #include "Src/DataStructure/FenwickTree/FenwickTree.hpp"
// #include "Src/DataStructure/SegmentTree/SegmentTree.hpp"
// #include "Src/DataStructure/DisjointSetUnion/DisjointSetUnion.hpp"




#include <concepts>

namespace zawa {

namespace concepts {

template <class T>
concept Semigroup = requires {
    typename T::Element;
    { T::operation(std::declval<typename T::Element>(), std::declval<typename T::Element>()) } -> std::same_as<typename T::Element>;
};

} // namespace concepts

} // namespace zawa


namespace zawa {

namespace concepts {

template <class T>
concept Identitiable = requires {
    typename T::Element;
    { T::identity() } -> std::same_as<typename T::Element>;
};

template <class T>
concept Monoid = Semigroup<T> and Identitiable<T>;

} // namespace

} // namespace zawa

namespace zawa {

namespace concepts {

template <class T>
concept MonoidWithAction = requires {
    requires Monoid<typename T::ValueMonoid>;
    requires Monoid<typename T::OperatorMonoid>;
    { T::mapping(
            std::declval<typename T::ValueMonoid::Element>(),
            std::declval<typename T::OperatorMonoid::Element>()
            ) } -> std::same_as<typename T::ValueMonoid::Element>; 
};

} // namespace concepts

} // namespace zawa

#include <bit>

namespace zawa {

template <concepts::MonoidWithAction S>
class LazySegmentTree {
public:

    using VM = S::ValueMonoid;

    using V = typename VM::Element;

    using OM = S::OperatorMonoid;

    using O = typename OM::Element;

    LazySegmentTree() = default;

    explicit LazySegmentTree(usize n) 
        : m_n{n}, m_sz{1u << (std::bit_width(n))}, m_dat(m_sz << 1, VM::identity()), m_lazy(m_sz << 1, OM::identity()) {}

    explicit LazySegmentTree(const std::vector<V>& a)
        : m_n{a.size()}, m_sz{1u << (std::bit_width(a.size()))}, m_dat(m_sz << 1, VM::identity()), m_lazy(m_sz << 1, OM::identity()) {
        std::ranges::copy(a, m_dat.begin() + inner_size());
        for (usize i = inner_size() ; --i ; ) recalc(i);
    }

    [[nodiscard]] inline usize size() const noexcept {
        return m_n;
    }

    [[nodiscard]] V operator[](usize i) {
        assert(i < size());
        return get(i, 1, 0, inner_size());
    }

    [[nodiscard]] V get(usize i) {
        return (*this)[i];
    }

    [[nodiscard]] V product(usize l, usize r) {
        assert(l <= r and r <= size());
        return product(l, r, 1, 0, inner_size());
    }

    void operation(usize l, usize r, const O& o) {
        assert(l <= r and r <= size());
        return operation(l, r, o, 1, 0, inner_size());
    }

    void assign(usize i, const V& v) {
        assert(i < size());
        assign(i, v, 1, 0, inner_size());
    }

    void operation(usize i, const O& o) {
        assert(i < size());
        operation(i, o, 1, 0, inner_size());
    }

private:

    using NodeInfo = std::tuple<usize, usize, usize>;

public:

    template <class F>
    requires std::predicate<F, V>
    usize maxRight(usize l, F f) {
        assert(l <= size());
        if (!f(VM::identity())) return l;
        if (l == size()) return size();
        std::vector<NodeInfo> ranges;
        partition_range(l, size(), ranges, 1, 0, inner_size());
        V prod = VM::identity();
        for (auto [nd, nl, nr] : ranges) {
            if (!f(VM::operation(prod, m_dat[nd]))) {
                return maxRight(f, prod, nd, nl, nr);
            }
            else {
                prod = VM::operation(prod, m_dat[nd]);
            }
        }
        return size();
    }

    template <class F>
    requires std::predicate<F, V>
    usize minLeft(usize r, F f) {
        assert(r <= size());
        if (!f(VM::identity())) return r;
        if (!r) return 0;
        std::vector<NodeInfo> ranges;
        partition_range(0, r, ranges, 1, 0, inner_size());
        V prod = VM::identity();
        for (auto [nd, nl, nr] : ranges | std::views::reverse) {
            if (!f(VM::operation(m_dat[nd], prod))) {
                return minLeft(f, prod, nd, nl, nr);
            }
            else {
                prod = VM::operation(prod, m_dat[nd]);
            }
        }
        return 0;
    }

private:

    usize m_n{}, m_sz{};

    std::vector<V> m_dat;

    std::vector<O> m_lazy;

    inline usize inner_size() const noexcept {
        return m_sz;
    }
    
    void recalc(usize nd) {
        // assert(nd < inner_size());
        m_dat[nd] = VM::operation(m_dat[nd << 1 | 0], m_dat[nd << 1 | 1]);
    }

    void propagate(usize nd) {
        // assert(nd < inner_size());
        for (usize ch : {nd << 1 | 0, nd << 1 | 1}) {
            m_dat[ch] = S::mapping(m_dat[ch], m_lazy[nd]);
            m_lazy[ch] = OM::operation(m_lazy[ch], m_lazy[nd]);
        }
        m_lazy[nd] = OM::identity();
    }

    V product(usize ql, usize qr, usize nd, usize nl, usize nr) {
        if (qr <= nl or nr <= ql) return VM::identity();
        if (ql <= nl and nr <= qr) return m_dat[nd];
        propagate(nd);
        const usize m = (nl + nr) >> 1;
        return VM::operation(
                product(ql, qr, nd << 1 | 0, nl, m),
                product(ql, qr, nd << 1 | 1, m, nr)
                );
    }

    V get(usize i, usize nd, usize nl, usize nr) {
        if (nd >= inner_size()) return m_dat[nd];
        propagate(nd);
        const usize m = (nl + nr) >> 1;
        return i < m ? get(i, nd << 1 | 0, nl, m) : get(i, nd << 1 | 1, m, nr);
    }

    void operation(usize ql, usize qr, const O& o, usize nd, usize nl, usize nr) {
        if (qr <= nl or nr <= ql) return;
        if (ql <= nl and nr <= qr) {
            m_dat[nd] = S::mapping(m_dat[nd], o);
            m_lazy[nd] = OM::operation(m_lazy[nd], o);
            return;
        }
        propagate(nd);
        const usize m = (nl + nr) >> 1;
        operation(ql, qr, o, nd << 1 | 0, nl, m);
        operation(ql, qr, o, nd << 1 | 1, m, nr);
        recalc(nd);
    }

    void operation(usize i, const O& o, usize nd, usize nl, usize nr) {
        if (nl == i and i + 1 == nr) {
            m_dat[nd] = S::mapping(m_dat[nd], o);
            // 葉頂点なので、lazyへのopは不要
            return;
        }
        propagate(nd); 
        const usize m = (nl + nr) >> 1;
        i < m ? operation(i, o, nd << 1 | 0, nl, m) : operation(i, o, nd << 1 | 1, m, nr);
        recalc(nd);
    }

    void assign(usize i, const V& v, usize nd, usize nl, usize nr) {
        if (nl == i and i + 1 == nr) {
            m_dat[nd] = v;
            return;
        }
        propagate(nd); 
        const usize m = (nl + nr) >> 1;
        i < m ? assign(i, v, nd << 1 | 0, nl, m) : assign(i, v, nd << 1 | 1, m, nr);
        recalc(nd);
    }

    void partition_range(usize ql, usize qr, std::vector<NodeInfo>& res, usize nd, usize nl, usize nr) {
        if (qr <= nl or nr <= ql) return;
        if (ql <= nl and nr <= qr) {
            res.emplace_back(nd, nl, nr);
            return;
        }
        propagate(nd);
        const usize m = (nl + nr) >> 1;
        partition_range(ql, qr, res, nd << 1 | 0, nl, m);
        partition_range(ql, qr, res, nd << 1 | 1, m, nr);
    }

    template <class F>
    requires std::predicate<F, V>
    usize maxRight(F f, const V& prod, usize nd, usize nl, usize nr) {
        if (nd >= inner_size()) return nl;
        propagate(nd);
        const usize m = (nl + nr) >> 1, lch = nd << 1 | 0, rch = nd << 1 | 1;
        return f(VM::operation(prod, m_dat[lch])) ? 
            maxRight(f, VM::operation(prod, m_dat[lch]), rch, m, nr) : maxRight(f, prod, lch, nl, m);
    }

    template <class F>
    requires std::predicate<F, V>
    usize minLeft(F f, const V& prod, usize nd, usize nl, usize nr) {
        if (nd >= inner_size()) return nr;
        propagate(nd);
        const usize m = (nl + nr) >> 1, lch = nd << 1 | 0, rch = nd << 1 | 1;
        return f(VM::operation(m_dat[rch], prod)) ? 
            minLeft(f, VM::operation(m_dat[rch], prod), lch, nl, m) : minLeft(f, prod, rch, m, nr);
    }
};

} // namespace zawa
using namespace zawa;
// #include "atcoder/modint"
// using mint = atcoder::modint998244353;
using namespace std;
#include <optional>
struct VM {
    using Element = optional<long long>;
    static Element identity() {
        return nullopt;
    }
    static Element operation(Element L, Element R) {
        if (!L) return R;
        if (!R) return L;
        return max(L.value(), R.value());
    }
};
struct OM {
    using Element = long long;
    static Element identity() {
        return 0LL;
    }
    static Element operation(Element L, Element R) {
        return L + R;
    }
};
struct ACT {
    using ValueMonoid = VM;
    using OperatorMonoid = OM;
    static VM::Element mapping(VM::Element v, OM::Element o) {
        if (!v) return o ? VM::Element{o} : nullopt; 
        else return v.value() + o;
    }
};
int N, A, X[200020], Y[200020], V[200020];
int main() {
    cin.tie(0);
    cout.tie(0);
    ios::sync_with_stdio(0);
    cin >> N >> A;
    A *= 2;
    for (int i = 0 ; i < N ; i++) {
        cin >> X[i] >> Y[i] >> V[i];
        X[i] *= 2;
        Y[i] *= 2;
    }
    CompressedSequence xs(vector<long long>(X, X + N)), ys(vector<long long>(Y, Y + N));
    vector<vector<int>> event(xs.size());
    for (int i = 0 ; i < N ; i++) event[xs.map(i)].push_back(i);
    LazySegmentTree<ACT> seg(ys.size());
    auto insert = [&](long long y, int v) -> void {
        seg.operation(ys[y - A - 1], ys.at(y) + 1, v); 
    };
    auto erase = [&](long long y, int v) -> void {
        seg.operation(ys[y - A - 1], ys.at(y) + 1, -v); 
    };
    auto eval = [&]() -> long long {
        auto prod = seg.product(0, seg.size());
        return prod ? prod.value() : 0LL;
    };
    int j = 0;
    long long ans = 0LL;
    for (int i = 0 ; i < ssize(event) ; i++) {
        const int x = xs.inverse(i);
        while (xs.inverse(j) + A < x) {
            for (int k : event[j]) erase(Y[k], V[k]);
            ans = max(ans, eval());
            j++;
        }
        for (int k : event[i]) insert(Y[k], V[k]); 
        ans = max(ans, eval());
    }
    for ( ; j < ssize(event) ; j++) {
        for (int k : event[j]) erase(Y[k], V[k]);
        ans = max(ans, eval());
        j++;
    }
    cout << ans << endl;
}