{
  "slug": "failure-detector",
  "term": "Failure detector",
  "aliases": [
    "distributed failure detector"
  ],
  "definition": "A failure detector is a component that uses observations to report whether another process or endpoint is suspected of being unavailable.",
  "note": "A suspicion can be mistaken and is not the same as resource fencing. The interpretation belongs to a particular observer and monitoring path. [hesela-phi]\n\nCassandra's documented UP/DOWN decisions are local; convicting an endpoint changes routing rather than automatically removing it from membership. Heartbeat observations alone do not establish application progress or durable write completion. [cassandra-membership]",
  "related": [
    "phi-accrual-detector",
    "gray-failure",
    "io-fencing",
    "io-timeout"
  ],
  "citations": [
    {
      "id": "cassandra-membership",
      "title": "Apache Cassandra documentation: Dynamo, Ring Membership and Failure Detection (reviewed 17 October 2026).",
      "url": "https://cassandra.apache.org/doc/latest/cassandra/architecture/dynamo.html#ring-membership-and-failure-detection",
      "accessed": "2026-10-17"
    },
    {
      "id": "hesela-phi",
      "title": "Phi failure detectors: suspicion is not a failure probability",
      "url": "https://hesela.com/analyses/phi-failure-detectors-suspicion/",
      "accessed": "2026-10-17"
    }
  ],
  "status": "shipping",
  "advantages": [],
  "limitations": [],
  "supersedes": [],
  "supersededBy": [],
  "firstSeen": null,
  "lastSeen": null,
  "updated": "2026-10-17"
}