Prior to rolling out on a large scale, we would like to be able to monitor the version of the particular beat installed. The beat exe's do not report any file version or product version. It is logged, but it can be programmatically a PITA to extract that line from the log, especially if the log has rolled before restart.
Could the file version be added to the next build?
Have you tried $ <beatname> -version ? You using rpm's ?
On can have a cron job writing these to some file every 24 hours for example. Use filebeat to push the versions.log file. For example <beatname> -version > version-$(date).log. Do not forget to delete old versions files from time to time.
I was unaware of the -version command. It would make a good last-ditch way of tracking if needed. We are primarily a windows shop, with only about 40 Linux boxes (almost 500 windows). We use BigFix for a great many things, and it is a simple matter to use BigFix to query a file for its version, and display it in a report.
We are okay for the moment; we have just started deployment, and are on the latest version. However, when the first upgrade comes down, we'll need to know which clients to push the update to.
This is why I was hoping that (at least for windows versions), the File Version property could be added to the next release.
as we don't have a version file yet, how about updating <beatname> -version > <path-to-data>/<beatname>/version to the install script. this way on every update you will have a valid version-file available all the time.
Apache, Apache Lucene, Apache Hadoop, Hadoop, HDFS and the yellow elephant
logo are trademarks of the
Apache Software Foundation
in the United States and/or other countries.