mcp-server-kafka
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| MCP_KAFKA_BOOTSTRAP_SERVERS | Yes | The Kafka bootstrap servers (e.g. localhost:9092) |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {
"listChanged": false
} |
| prompts | {
"listChanged": false
} |
| resources | {
"subscribe": false,
"listChanged": false
} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| kafka_list_connectionsA | 枚举当前服务端已配置的所有 Kafka 集群命名连接清单、Broker 引导地址与只读保护状态 |
| kafka_cluster_infoB | 获取 Kafka 集群整体元数据与 Broker 节点列表（包含 Cluster ID、控制器与节点地址） |
| kafka_list_topicsC | 列出 Kafka 集群中的主题清单（支持按名称模式过滤并默认排除内部系统主题） |
| kafka_describe_topicB | 查询指定 Kafka 主题的详细拓扑（分区分布、Leader 节点、ISR 同步副本与配置） |
| kafka_sample_messagesA | 从指定 Kafka 主题以零提交位移方式只读采样消息（支持 latest/earliest/offset 策略、自适应解码与大报文截断保护） |
| kafka_list_consumer_groupsB | 列出 Kafka 集群中的消费组清单（包含 Group ID、协议类型与活跃状态如 Stable/Empty/Dead） |
| kafka_describe_consumer_groupB | 查询指定 Kafka 消费组的详细拓扑，包含活跃成员、分区 Committed Offset、LEO 及 Lag 积压数值 |
| kafka_create_topicA | 在 Kafka 集群中创建新主题（支持指定分区数与副本因子；只读模式下不可用） |
| kafka_delete_topicA | 从 Kafka 集群中删除主题（高危破坏操作，必须传入 confirm=True 显式确认） |
| kafka_produce_messageB | 向指定 Kafka 主题发送消息（支持 Key、JSON/文本载荷与 Headers；只读模式下不可用） |
| kafka_reset_consumer_group_offsetsA | 重置指定消费组在目标主题上的消费位移（高危运维操作，支持 Dry-Run 预检与 earliest/latest/to_offset/to_datetime 策略） |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
| diagnose_topic_lag | 引导大模型对 Kafka 消费组滞后 (Lag) 进行深度根因分析与拓扑倾斜诊断 |
| inspect_topic_messages | 引导大模型对 Kafka 主题消息数据进行采样探查、载荷反序列化校验与业务数据清洗分析 |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
| Kafka 集群拓扑摘要 | 以只读上下文形式输出当前 Kafka 集群整体元数据与 Broker 运行节点摘要 |
TDQS
Scored across 11 tools
Each tool targets a distinct Kafka resource and action: list/describe topics, sample messages, list/describe consumer groups, create/delete topic, produce message, list connections, cluster info, and reset offsets. The list-vs-describe pattern is standard and descriptions clarify scope. No two tools are likely to be confused for the same operation.
All names use snake_case with a kafka_ prefix, and most follow verb_noun (list_topics, describe_topic, create_topic). One outlier is kafka_cluster_info, which is noun_noun rather than verb_noun, and singular/plural varies slightly (topic vs topics). Still largely predictable.
11 tools is well within the sweet spot for a Kafka admin/ops server, covering both read and write/destructive operations without bloat. Each tool maps to a clear capability; no redundant or filler tools.
The surface covers core Kafka lifecycle: cluster info, topic list/describe/create/delete, message sample/produce, consumer group list/describe/reset offsets. Minor gaps remain, such as altering topic partitions/configs or deleting consumer groups, but core workflows are covered. Agents can work around most gaps using existing tools.