This looks exactly like a CPU instruction set mismatch issue rather than a service configuration problem.
If the VM host doesn’t support AVX/AVX2, newer Elasticsearch/Kibana builds will fail to start because they’re compiled with higher -march targets.
At this point, the quickest fix is usually either:
-
confirm CPU flags on the host (
lscpu | grep avx) -
downgrade to an Elasticsearch/Kibana version built for older architectures, or
-
move the deployment to a VM that supports AVX/AVX2
Rebuilding with a lower -march can work too, but in most cases upstream binaries are the real constraint here, so version alignment is usually the cleaner solution.