ReportPortal analyzer服务无法连接RabbitMQ问题求助

问题现象 #

通过K8s部署ReportPortal（Helm Chart版本25.5.30），整体运行正常，但自动分析功能失效，核心原因是analyzer服务无法启动。

相关日志 #

API Pod日志 #

2025-06-03 15:32:47.705 2025-06-03 13:32:47,609 - [un8qThWMTjKjDSy8SF4-4Q] - INFO - analyzerApp.amqp - Trying to connect to amqp://reportportal-rabbitmq.reportportal.svc.cluster.local:5672/analyzer?heartbeat=30
2025-06-03 15:32:47.706 2025-06-03 13:32:47,609 - [xy61hSzYS9-D13MzhYNnNA] - INFO - analyzerApp.amqp - AMQP connection established.
2025-06-03 15:32:47.706 2025-06-03 13:32:47,604 - [ofVyUEV-QzapYgMIoePE7w] - ERROR - analyzerApp.amqp - Connection/channel lost. Reconnecting. Exchange: 'analyzer-default'. Queue: 'index_suggest_info'.
2025-06-03 15:32:47.706 Traceback (most recent call last):
2025-06-03 15:32:47.706   File "/backend/app/amqp/amqp.py", line 187, in receive
2025-06-03 15:32:47.706     self._bind_queue(channel, queue, self._config.amqpExchangeName)
2025-06-03 15:32:47.706   File "/backend/app/amqp/amqp.py", line 147, in _bind_queue
2025-06-03 15:32:47.706     channel.queue_bind(exchange=exchange_name, queue=name, routing_key=name)
2025-06-03 15:32:47.706   File "/venv/lib64/python3.11/site-packages/pika/adapters/blocking_connection.py", line 2570, in queue_bind
2025-06-03 15:32:47.706     self._flush_output(bind_ok_result.is_ready)
2025-06-03 15:32:47.706   File "/venv/lib64/python3.11/site-packages/pika/adapters/blocking_connection.py", line 1339, in _flush_output
2025-06-03 15:32:47.706     raise self._closing_reason  # pylint: disable=E0702
2025-06-03 15:32:47.706     ^^^^^^^^^^^^^^^^^^^^^^^^^^
2025-06-03 15:32:47.706 pika.exceptions.ChannelClosedByBroker: (404, "NOT_FOUND - no exchange 'analyzer-default' in vhost 'analyzer'")

RabbitMQ Pod日志 #

2025-06-03 16:00:36.600 2025-06-03 14:00:36.117476+00:00 [error] <0.8690128.0> Channel error on connection <0.8690983.0> (10.19.172.218:43084 -> 10.19.227.194:5672, vhost: 'analyzer', user: 'rabbitmq'), channel 1:
2025-06-03 16:00:36.600 2025-06-03 14:00:36.117476+00:00 [error] <0.8690128.0> operation queue.bind caused a channel exception not_found: no exchange 'analyzer-default' in vhost 'analyzer'

排查信息 #

已确认analyzer-default交换机未被创建，未修改analyzer服务的任何默认配置。

1. 检查RabbitMQ初始化Job状态 #

ReportPortal部署时，RabbitMQ的交换机、队列等资源由rabbitmq-init-job负责初始化。先确认该Job是否成功执行：

# 查看Job状态
kubectl get jobs -n reportportal
# 查看Job执行日志
kubectl logs <rabbitmq-init-job-pod-name> -n reportportal

如果Job执行失败，根据日志修复问题后，重新执行该Job：

kubectl delete job <rabbitmq-init-job-name> -n reportportal
# 等待Helm自动重建Job或手动触发

2. 手动创建缺失的交换机 #

若初始化Job无法正常修复，可直接在RabbitMQ的analyzer虚拟主机中创建analyzer-default交换机：

方式一：通过RabbitMQ管理界面

• 访问RabbitMQ管理控制台，切换到analyzer虚拟主机
• 进入Exchanges页面，点击Add a new exchange
• 填写名称为analyzer-default，类型选择topic，勾选Durable，点击Add exchange

方式二：通过命令行操作

# 进入RabbitMQ Pod
kubectl exec -it <rabbitmq-pod-name> -n reportportal -- bash
# 授予用户权限（若未授权）
rabbitmqctl set_permissions -p analyzer rabbitmq ".*" ".*" ".*"
# 创建交换机
rabbitmqadmin declare exchange name=analyzer-default type=topic durable=true vhost=analyzer

3. 重启analyzer服务 #

交换机创建完成后，重启analyzer Pod：

kubectl rollout restart deployment/reportportal-analyzer -n reportportal

4. 验证功能恢复 #

等待analyzer Pod启动完成后，查看Pod日志确认无AMQP连接错误，同时测试ReportPortal的自动分析功能是否正常运行。

内容的提问来源于stack exchange，提问作者makeiteasy