{"version":"2.1.238","anchor":"mcp-handshake-timeouts-are-recorded-as-lost-connections","canonical_anchor":"mcp-handshake-timeouts-are-recorded-as-lost-connections","heading":"MCP handshake timeouts are recorded as lost connections","tier":"internal","area":"MCP","scope":null,"heads_up":null,"url":"https:\/\/changelogs.core-directive.com\/v\/2.1.238\/e\/mcp-handshake-timeouts-are-recorded-as-lost-connections","release_url":"https:\/\/changelogs.core-directive.com\/v\/2.1.238","markdown":"### MCP handshake timeouts are recorded as lost connections\n\nAn MCP handshake that times out is now recorded as a lost connection instead of an ordinary shutdown.\n\n**What**\n\nWhen an MCP handshake times out or a stale client row is cleaned up, the task is now torn down as a lost connection rather than an ordinary quiet kill, so the two causes can be told apart downstream.\n\n**Evidence**\n\n`connectionLost: !0`\n\n- Area: MCP\n- Tier: Under the hood\n- Useful: 2\/5\n- Signal: 2\/5"}