I second the remark from @RomeoNinov that arguments for/against either tool are more likely to generate opinions than facts. Too long for a comment:
relative market shares unknown
With opensource tools market share data based on for example published sales figures are missing.
Unlike for example opensource web server software, management tooling is almost always not publicly accessible and shielded from direct internet access. That makes collecting hard data about deployment counts in the wild also nigh impossible.
knowledge of one tool seldom a hard requirement
A fresh graduate will not be hired for their extensive knowledge of either Zabbix or Nagios, nor is it likely that they'll be asked to set up either a Zabbix or Nagios monitoring and alert system from scratch either. They won't be hired as Zabbix or Nagios administrator. IMHO
They'll be hired to join an existing team as for example junior network or system administrator.
There monitoring will only be (a small) part of their duties and tasks and in-depth knowledge of the same specific monitoring tool that our team happens to use already, that has never been a hard requirement for us and is at best a nice to have. Knowledge of the specific tool we do use can be learned on the job.
What is important to us is the translatable skillset: knowledge of a monitoring tool, an idea of what observability means in the real world, why one would set certain thresh-holds, when and why you'd need an agent, how to do trend analyses, visualisation and how to draw meaningful conclusions etc. etc.
