Kibana always crashes

Every day or so, I noticed that my instance of kibana on AWS AMI disappears. It must have crashed.

I have to log back into kibana to restart it.

I look at the /var/log/kibana/ and see two log files

kibana.stdout
kibana.stderr

kibana.stdout does not show any info that gives any indication of a error or severe situation.

kibana.stderr shows this:

FATAL { [Error: EACCES, open '/opt/kibana/optimize/bundles/kibana.bundle.js']
cause:
{ [Error: EACCES, open '/opt/kibana/optimize/bundles/kibana.bundle.js']
errno: -13,
code: 'EACCES',
path: '/opt/kibana/optimize/bundles/kibana.bundle.js' },
isOperational: true,
errno: -13,
code: 'EACCES',
path: '/opt/kibana/optimize/bundles/kibana.bundle.js' }

I looked at the read/write permission of that path. It is read and write accessible for any one.

On AMI AWS, I think I have kibana 4.4 installed:

I used this instruction to install it on AMI

https://www.elastic.co/guide/en/kibana/current/setup.html

And my /etc/yum.repos.d/kibana.repo contains this:

[kibana-4.4]
name=Kibana repository for 4.4.x packages
baseurl=http://packages.elastic.co/kibana/4.4/centos
gpgcheck=1
gpgkey=http://packages.elastic.co/GPG-KEY-elasticsearch
enabled=1

Are there timestamps associated with those error messages?

Can you check the timestamp of that kibana.bundle.js?

I was told by a developer that the file is only written to when a plugin is installed or when some other kibana file is manually changed. Anything like that happening?