Startup crash due to failure to bind IPv6 address in IPv6 EKS cluster in Fluentbit 5.x.x

[paolofacchinetti]

Bug Report

Describe the bug
After upgrading fluentbit from version 4.1.1 to 5.0.2, the new fluentbit pods crash briefly after startup due to this error:

[2026/04/03 12:16:25.735] [ warn] net_server: getaddrinfo(listen='[::]:2020'): Name or service not known
[2026/04/03 12:16:25.735] [error] [downstream] could not bind address [::]:2020. Aborting

I have deployed fluent-bit via the official Helm Chart, in an IPv6 based EKS cluster. Needless to say it worked perfectly before. In the 5.x.x release notes there's no mention of any breaking change related to IPv6 environments.

Have also noticed another user complaining about this problem in another issue: #11620 (comment)

** Full Helm configuration used **

Posting the "service" section of the values.yaml I'm using to deploy fluentbit, as mentioned earlier it worked perfectly fine before and I have not changed anything when deploying fluentbit 5.x.x

# -- FluentBit config
config:
  service: |
    [SERVICE]
        Daemon Off
        Flush {{ .Values.flush }}
        Log_Level {{ .Values.logLevel }}
        Parsers_File /fluent-bit/etc/parsers.conf
        Parsers_File /fluent-bit/etc/conf/custom_parsers.conf
        HTTP_Server On
        HTTP_Listen [::]
        HTTP_Port {{ .Values.metricsPort }}
        Health_Check On

Your Environment

• Version used: 5.0.2 (helm chart 0.57.2)
• Configuration: see above
• Environment name and version (e.g. Kubernetes? What version?): EKS 1.35
• Server type and version: AmazonLinux ec2, mixed arm64 and amd64