Most of my free engineering time goes into open source. Almost every project here started as a fix for my own problem at work. Then other people had the same problem. In my opinion, that is the only good reason to start an open source project.
Found a bug? Open an issue. Can you fix it? Even better, open a pull request.
Helm plugin that decrypts sops or vals encrypted values files on the fly. Keep secrets in Git, next to your charts. Works with Helm, Helmfile and Argo CD.
Secrets in Git are fine when they are encrypted. Secrets in a wiki page are not.
Single sign-on for the OpenVPN server. Users log in with Entra ID, Keycloak or any other OIDC provider, and you get MFA and conditional access for free.
Static VPN passwords should have died years ago.
The Prometheus exporter for Windows machines. Think node_exporter, but for Windows hosts and Windows nodes in Kubernetes.
Windows nodes still exist. They need monitoring, too.
Receives nginx access logs over syslog and turns them into Prometheus metrics: request rates, status codes and latencies.
You do not need a log pipeline to count HTTP 500s.
Lens extension that shows the expiry date of certificates stored in Kubernetes secrets.
Expired certificates are the most boring outage there is.
I’m the general admin of the prometheus-community and grafana-community Helm chart repositories, and I maintain charts there and in Sveltos. A widely used chart is an API: every renamed value breaks someone’s production. I review accordingly.
General admin of the whole repository. Charts I maintain directly:
kube-prometheus-stack
Prometheus Operator, Prometheus, Alertmanager, Grafana and default rules in one chart.prometheus-operator-crds
The Prometheus Operator CRDs, with their own lifecycle.prometheus-windows-exporter
windows_exporter as a DaemonSet for Windows nodes.prom-label-proxy
Label-based access control for Prometheus, Thanos and Alertmanager APIs.jiralert
Jira tickets from Alertmanager notifications.General admin of the whole repository, including CI and Renovate. Charts I maintain directly:
grafana
The Grafana chart.loki
The Loki chart.grafana-mcp
The Grafana MCP server.projectsveltos
Sveltos, add-on management for tens of Kubernetes clusters.More on my GitHub profile.