Observability
Two collectors feed two stores, and Grafana reads both. Alertmanager ships inside the VictoriaMetrics stack, not as its own app.
flowchart LR pods["Container logs<br/>/var/log/pods"] --> agent["OTel agent<br/>DaemonSet"] falco["Falco<br/>JSON to stdout"] --> pods kubelet["kubelet + host metrics"] --> agent api["Kubernetes API<br/>cluster metrics + events"] --> gw["OTel gateway<br/>Deployment"] sm["ServiceMonitors<br/>apps, DCGM, Argo CD"] --> vmagent["VMAgent"] agent -->|remote write| vm["VictoriaMetrics<br/>VMSingle"] gw -->|remote write| vm vmagent --> vm agent -->|OTLP/HTTP| vl["VictoriaLogs"] gw -->|OTLP/HTTP| vl vm --> vmalert["VMAlert → Alertmanager"] vm --> grafana["Grafana"] vl --> grafana
VictoriaMetricsVMSingle, VMAgent, VMAlert and Alertmanager from one chart.
VictoriaLogsLog store fed over OTLP, queried through Grafana's Loki plugin.
OpenTelemetryAgent DaemonSet and gateway Deployment collector pipelines.
Argo CD monitorMetrics Services, ServiceMonitors and the TLS certificate for Argo CD.