Skip to content
Mainframe

B2B — Automation & Build

Grafana Monitoring Stack

Self-hosted Prometheus + Grafana + Loki on your infrastructure — dashboards, alerting, and onboarding.

A full self-hosted observability stack for your applications and infrastructure — no Datadog bill, no data flowing to third parties, with dashboards you’ll actually look at.

What’s included

  • Prometheus, Grafana, and Loki running on your server (Hetzner VM, Docker, or Kubernetes)
  • Data source configuration: application metrics, Node Exporter for server metrics, log aggregation from containers and systemd units
  • Dashboards in three layers:
    • Application health (latency, error rate, saturation — the four golden signals)
    • Infrastructure (CPU, memory, disk, network per host)
    • Business KPIs agreed with you (orders per hour, conversion rate, active sessions)
  • Alerting via Alertmanager: email, Slack, or PagerDuty — with sensible thresholds, not “80% on everything”
  • Backup strategy for configuration and long-term metrics
  • Onboarding to the platform — scope agreed individually: how to read the dashboards, how to interpret an alert, how to build a new panel

Who it’s for

SMBs running applications in production who need clear visibility into performance and availability without locking into a SaaS provider. Also a good fit for migrating off Datadog, New Relic, or similar.

What’s not included

No application instrumentation in code (e.g. tracing integration into your services) — that’s a separate engagement, depending on the stack. No 24/7 on-call service.