yarn Hadoop airflow mysql dag operator ops troubleshooting Ambari exception Flink hdfs others Kafka atlas ranger Spark Hive flume dolphinScheduler kerberos ClickHouse olap tableEngine metaData clickhouseFunctions Greenplum mapreduce sqoop datanode HBase hiveFunctions hudi data-lake lakehouse timeline instant index cow mor write upsert spark sql performance flink-sql catalog cdc JVM other idea avro langchain langgraph langsmith qwen mlx openai message llm prompt output-parser structured-output chatbot memory gradio mlx-whisper multimodal mlx-vlm agent workflow cli studio api sdk tool streaming context configurable agentstate postgresql edge asyncio evaluator mcp human-in-the-loop interrupt command state reducer json-rpc fastmcp python jwt oauth storm security hexo centos linux cloudstack Zookeeper fastdfs azkaban Iterm2 yum rpm markdown 爬虫 Phoenix rag embeddings cosine-similarity qwen3 sentence-transformers vector-database semantic-search faiss vector-search chroma milvus retriever sqlite redis datatype cache nosql rdb aof transaction pipeline replica replication sentinel high-availability cluster springboot io epoll bigkey canal bitmap hyperloglog bloom-filter availability lock lua redisson redlock source-code data-structure Scala Maven sparkCore sparkStreaming spark streaming sparkSQL sparkMl spark ml parquet prometheus monitoring grafana