Kibana and ES version: 7.13.2
I am running cluster of 5 nodes (3 ES data & master nodes and 2 coordinating nodes) and one Kibana. Whole cluster is running in docker swarm on the same network. My ES cluster works fine, I can connect to it with Cerebro or Kibana from localhost. When I tried connect to ES from Kibana which is deployed to swarm, it ends up with:
2021-07-08T10:39:16.963979391Z {"type":"log","@timestamp":"2021-07-08T11:39:16+01:00","tags":["info","savedobjects-service"],"pid":953,"message":"Waiting until all Elasticsearch nodes are compatible with Kibana before starting saved objects migrations..."}
2021-07-08T10:39:17.031184806Z {"type":"log","@timestamp":"2021-07-08T11:39:17+01:00","tags":["error","savedobjects-service"],"pid":953,"message":"Unable to retrieve version information from Elasticsearch nodes."}
I set ELASTICSEARCH_URL variable in Kibana and I am able to reach ES cluster from Kibana container with curl as:
- coordinating nodes service name - OK
- container IP - OK
- network alias - OK
- container name - OK
When I reach Kibana from browser, I can see:
Kibana server is not ready yet
I can see no kibana indices in ES.
When I run Kibana with command as docker run ...
I am able to connect to ES cluster.
Questions:
- Why I cannot connect to ES with Kibana from swarm but from localhost I can?
- What is savedobjects-service? I have experience with older versions of ELK stack and do not remember logs like this.