After installing fluentbit package into a VKS cluster for VCF Operations monitoring, there are millions of blank, empty events in VCF Operations.
Fluentbit has been set up with the steps and sample YAML from the vSphere Supervisor Techdocs below:
When viewing fluentbit pod logs in the VKS workload cluster, the following errors are present:
The fluentbit IP and port are the same values configured in the fluent-bit-data-values.yaml
kubectl get pods -n tkg-system
kubectl logs -n tkg-system <fluent-bit pod>
YYYY-MM-DDTHH:MM:SS.ssssssZ stderr F [2025/11/09 08:54:26] [error] [net] TCP connection failed: <fluentbit IP>:<port> (No route to host)
YYYY-MM-DDTHH:MM:SS.ssssssZ stderr F [2025/11/09 08:54:26] [error] [output:syslog:syslog.0] no upstream connections available
YYYY-MM-DDTHH:MM:SS.ssssssZ stderr F [2025/11/09 08:54:29] [error] [upstream] connection #47 to tcp://<fluentbit IP>:<port> timed out after 10 seconds (connection timeout)
VCF Operations 9 or higher
vSphere Supervisor
Ensure that the correct IP and port is configured in the fluent-bit-data-values.yaml.
The outputs section must have a properly set Host and Port for VCF Operations and Log Insight according to the desired Connection Type:
Values in brackets <> should be replaced according to your configuration needs and environment.
outputs: |
[OUTPUT]
Name <connection type>
Match *
Host <fluentbit IP>
Port <fluentbit port>
URI api/v2/events
Format json
tls <on/off>
tls.debug 4
tls.verify off
json_date_key timestamp
See the below table for Ports organized by Connection Type:
| Connection Type | Port | Protocol |
| syslog | 514 | UDP |
| syslog | 6514 | TCP |
| http | 9000 | |
| https | 9543 | |
| tls on | 9543 |
If the fluent-bit-data-values.yaml needs to be updated:
vcf packages installed list -A
vcf packages installed update -n <fluentbit package namespace> <fluentbit package name> --values-file <fluent-bit-data-values.yaml>
kubectl get pods -A | grep fluentbit
kubectl rollout restart daemonset fluentbit -n <fluentbit namespace>
If there are concerns that there are networking issues, you can create a custom, simple pod with curl in the same node as a fluentbit pod and confirm that the fluentbit IP can be reached.
Once the custom, simple pod with curl is created and running in the workload cluster node, the below command can be used to curl the fluentbit IP and port above:
kubectl exec -it -n <custom pod namespace> <custom pod name> -- curl -vk <fluentbit IP>:<fluentbit port>