Kubernetes Services: Filtering, Dependency Chain Traversal, and Status Query

Coding: Kubernetes Services Filtering + Dependency Chain Traversal + Status Query You are given a collection of Kubernetes service descriptors. Each service contains at least: name: unique service name (string) attributes: key/value metadata (string -> string) status: one of Healthy / Degraded / Unhealthy / Unknown dependencies: list of service names this service depends on (directed edge service -> dependency) Implement three parts: Part 1: Filtering Implement filtering over the service list and return names (preserve input order) that satisfy all provided conditions (AND): status_in: allowed set of statuses attribute_equals: require a given attribute key to exist and equal a given value name_prefix: match a name prefix Part 2: Dependency chain + aggregated status Given start_service, return: reachable: all nodes reachable from start_service following dependencies (including start), in BFS visitation order aggregated_status: aggregated status over the reachable set: If any node is Unhealthy -> Unhealthy Else if any is Degraded -> Degraded Else if all are Healthy -> Healthy Otherwise -> Unknown Handle missing dependency references (ignore) and cycles (do not loop forever). Part 3: Open-ended design If you wrap this into an MCP server for an LLM Agent: What tools/APIs would you expose? How would you reduce noise and prevent misuse (validation, paging, top-k, sampling, thresholds, dedup, idempotency, rate limiting, caching, authz, etc.)? Constraints N services: 1..200000 E dependencies: 0..500000 Part2 may be called many times I/O (for an online coding environment) Assume JSON Lines: Line 1: JSON array of services Line 2: JSON object for filter conditions Line 3: JSON string for start_service Output: Line 1: JSON array of filtered service names Line 2: JSON object: { "reachable": [...], "aggregated_status": "..." } Example Input [{"name":"api","attributes":{"team":"a"},"status":"Healthy","dependencies":["db","cache"]},{"name":"db","attributes":{"team":"a"},"status":"Degraded","dependencies":[]},{"name":"cache","attributes":{"team":"b"},"status":"Healthy","dependencies":[]}] {"status_in":["Healthy"],"attribute_equals":{"key":"team","value":"a"},"name_prefix":"a"} "api" Output ["api"] {"reachable":["api","db","cache"],"aggregated_status":"Degraded"}