Falco writes each event as one line of JSON. They end up in Loki with the rest of the pod logs, and the counts per rule end up in Prometheus.
#In Grafana
Forward Grafana, open Explore and pick Loki:
kubectl -n application port-forward svc/grafana 3000:3000All Falco events:
{namespace="falco"} | json | rule != ""Only critical ones:
{namespace="falco"} | json | priority="Critical"Events from one namespace’s pods:
{namespace="falco"} | json | output_fields_k8s_ns_name="cluster-1"Each event includes the rule name, the priority, the command line, the container and the Kubernetes pod and namespace.
#From the pods
To read the latest events on every node without Grafana:
kubectl -n falco logs -l app=falco --tail=50 --prefix#Which rules fire most
In Prometheus, at http://localhost:9090 after forwarding svc/prometheus-operated:
sort_desc(sum by (rule_name) (increase(falcosecurity_falco_rules_matches_total[24h])))This shows which rules matched in the last day and how often. A rule that fires constantly on a known component, such as Longhorn or CloudNativePG doing normal work, is a candidate for an exception in Falco’s rules.