Describe the bug
I ran into the below issue in 3.2.0 version that cpu usages could be negative and the same value (-1)
$ curl "https://dev-opensearch-test-250821-os.osp01.abc.com/_cat/nodes?v" -ku admin:test
ip heap.percent ram.percent cpu load_1m load_5m load_15m node.role node.roles cluster_manager name
10.251.175.203 38 60 -1 0.44 0.41 0.49 m cluster_manager * dev-opensearch-test-250821-master-1
10.251.117.144 20 60 -1 0.58 0.54 0.60 m cluster_manager - dev-opensearch-test-250821-master-0
10.251.225.190 67 90 -1 1.79 1.14 1.03 di data,ingest - dev-opensearch-test-250821-data-0
10.251.143.196 24 60 -1 0.40 0.68 0.61 m cluster_manager - dev-opensearch-test-250821-master-2
Related component
Clients
To Reproduce
- deploy opensearch cluster using 3.2.0 version
- After your cluster runs successfully,
curl "https://dev-opensearch-test-250821-os.osp01.abc.com/_cat/nodes?v" -ku admin:test
- It also happens wrongly when you try
_cluster/stats API
Expected behavior
.
Additional Details
Plugins
I just added the prometheus-exporter plugin by building new Docker image
FROM harbor.abc.com/library/opensearchproject/opensearch:3.2.0
COPY prometheus-exporter-3.2.0.0.zip /tmp/
RUN ./bin/opensearch-plugin install file:/tmp/prometheus-exporter-3.2.0.0.zip
Host/Environment (please complete the following information):
- OS: All hosts are worker nodes of kubernetes v1.31.4 and based on RHEL 8.10 (4.18.0 kernel)
- Version: OpenSearch 3.2.0
Additional context
- Air-gapped environment but I believe it doesn't matter with this issue.
Describe the bug
I ran into the below issue in 3.2.0 version that cpu usages could be negative and the same value (-1)
Related component
Clients
To Reproduce
curl "https://dev-opensearch-test-250821-os.osp01.abc.com/_cat/nodes?v" -ku admin:test_cluster/statsAPIExpected behavior
.
Additional Details
Plugins
I just added the prometheus-exporter plugin by building new Docker image
Host/Environment (please complete the following information):
Additional context