René's URL Explorer Experiment

{"@context":"https://schema.org","@type":"DiscussionForumPosting","headline":"[BUG] KeepAliveScheduler does not evict sessions after ping failure, causing unbounded CPU growth","articleBody":"## Description\n\nWhen `keep-alive-interval` is configured on a Streamable HTTP MCP Server, `KeepAliveScheduler` periodically sends ping messages to all registered `McpStreamableServerSession` instances. When a session's stream is unavailable (client already disconnected), the ping fails and a WARN is logged, but **the session is never removed from the session map**. Over time, dead sessions accumulate indefinitely, causing increasing CPU usage as the scheduler iterates over a growing list of unreachable sessions every interval.\n\n## Environment\n\n- Spring AI: 1.1.4\n- MCP Java SDK: (bundled with Spring AI 1.1.4)\n- Java: JDK 25\n- Server: Tomcat (embedded via Spring Boot 3.4.1)\n- Transport: Streamable HTTP (`spring.ai.mcp.server.protocol=STREAMABLE`)\n\n## Configuration\n\n```yaml\nspring:\n  ai:\n    mcp:\n      server:\n        type: SYNC\n        protocol: STREAMABLE\n        streamable-http:\n          mcp-endpoint: /mcp\n          keep-alive-interval: 30s\n```\n\n## Steps to Reproduce\n\n1. Deploy an MCP Server with Streamable HTTP transport and `keep-alive-interval: 30s`\n2. Have multiple MCP clients connect, call tools, then disconnect (close TCP)\n3. Wait a few minutes and observe logs and CPU usage\n\n## Observed Behavior\n\nLogs fill with repeated warnings every 30 seconds, one per dead session:\n\n```\nWARN [io.modelcontextprotocol.util.KeepAliveScheduler]\nFailed to send keep-alive ping to session McpStreamableServerSession@3656e950: Stream unavailable for session d0a2a5b9-d427-49f6-8f0b-6c37293e2e35\n\nWARN [io.modelcontextprotocol.util.KeepAliveScheduler]\nFailed to send keep-alive ping to session McpStreamableServerSession@917a226: Stream unavailable for session 5e87f163-f127-4793-9fec-cd35b6a59327\n\nWARN [io.modelcontextprotocol.util.KeepAliveScheduler]\nFailed to send keep-alive ping to session McpStreamableServerSession@5795b3a2: Stream unavailable for session ...\n```\n\n- Dead sessions are **never** removed from the internal map\n- CPU usage grows linearly with the number of accumulated dead sessions\n- The `boundedElastic` thread pool is consumed by failed ping attempts\n- Eventually causes CPU alerts and degrades ability to serve new requests\n\nIn our production environment, CPU went from normal to sustained **135%+** within hours of deployment, triggered by ~50 disconnected clients accumulating dead sessions.\n\n## Expected Behavior\n\nWhen a ping fails (stream unavailable), `KeepAliveScheduler` should:\n\n1. Increment a failure counter for that session\n2. After N consecutive failures (e.g., 2-3), call `session.close()` and remove it from the session map\n3. Log the eviction at INFO level\n\n## Impact\n\n- **Severity: High** — causes progressive CPU degradation in production\n- Combined with the CLOSE-WAIT socket leak issue, can render a server completely unusable\n- No recovery without restart — dead sessions accumulate forever\n\n## Workaround\n\nDisable keep-alive entirely:\n\n```yaml\nspring:\n  ai:\n    mcp:\n      server:\n        streamable-http:\n          keep-alive-interval: 0s\n```\n\nThis stops the CPU waste but also removes the ability to detect stale sessions proactively.\n\n## Suggested Fix\n\nIn `KeepAliveScheduler`, after a failed ping:\n\n```java\n// Pseudocode\nvoid sendKeepAlive() {\n    Iterator\u003cMap.Entry\u003cString, McpStreamableServerSession\u003e\u003e it = sessions.entrySet().iterator();\n    while (it.hasNext()) {\n        Map.Entry\u003cString, McpStreamableServerSession\u003e entry = it.next();\n        try {\n            entry.getValue().sendPing();\n            failureCounters.remove(entry.getKey());\n        } catch (Exception e) {\n            int failures = failureCounters.merge(entry.getKey(), 1, Integer::sum);\n            if (failures \u003e= MAX_FAILURES) {\n                entry.getValue().close();\n                it.remove();\n                failureCounters.remove(entry.getKey());\n                log.info(\"Evicted dead session: {}\", entry.getKey());\n            }\n        }\n    }\n}\n```\n","author":{"url":"https://github.com/lxq19991111","@type":"Person","name":"lxq19991111"},"datePublished":"2026-06-11T04:24:42.000Z","interactionStatistic":{"@type":"InteractionCounter","interactionType":"https://schema.org/CommentAction","userInteractionCount":6},"url":"https://github.com/1022/java-sdk/issues/1022"}