ELK: NGINX logs analyze

Hello to everyone,
i have an strange issue. I'm running 2 nginx balancing units and it export it logs via filebeat to logstash and to elasticsearch index as a final destination.
In a same time im running nignx stats modules on this appliances. They also exports same stats to prometheus exporter and to my zabbix.
My question is why there is such a big difference on data which i'm getting in Kibana and Zabbix/Prometheus?
In Kibana i see peaks like 6000+ requests (hits) in a time window and about 300-400 requests in Zabbix. Nginx stats module also showing data like Zabbix showing me when i curl it.
Whats wrong with it?

Welcome to our community! :smiley:

To be clear, are you saying Filebeat is sending to Elasticsearch as well as zabbix and prometheus?