A few questions to help me better understand what may be happening in your situation:
Are you running Kibana behind a proxy? If so, are there any error logs in the proxy's logs?
Are other functions working correctly? Can you view dashboards or visualizations? If so, it is only failing when you try to save things?
What is the response of hitting the /_status endpoint in Elasticsearch? It's possible you have a shard-related issue in the cluster.
(I have saved almost 165 objects in the kibana,Is it possible that there are too many saved objects.)
This should not be a problem unless you have an Elasticsearch cluster that does not have enough disk space or available RAM. We support Kibana deployments with 1000s of objects.
second, other functions work correctly, I can view dashboards or visualizations. it is only failing(show "Service Unavailable") when I try to save things or create a index pattern. However, after refreshing several times, the visualization that failed to save will eventually appear in the list.
third, elasicsearch's status is green, I do not find the problem.
This is a strange one indeed. I haven't seen this type of behavior in Kibana before (503s only on POST requests).
However, after refreshing several times, the visualization that failed to save will eventually appear in the list
This is even stranger yet and points to a potential caching issue.
You mentioned that there are no proxies between your browser and the Kibana server. Is this server running bare on a machine you are accessing directly? Is it possible requests are being routed through nginx or a service mesh layer like Kubernetes?
Additionally, could you take a peek at the failing request in the "Network" tab of the dev tools? Specifically it'd be helpful to see the "Response headers" and the response body (available in the "Preview" tab of the request details pane).
Apache, Apache Lucene, Apache Hadoop, Hadoop, HDFS and the yellow elephant
logo are trademarks of the
Apache Software Foundation
in the United States and/or other countries.