Central Logging / Monitoring at VSHN

These pages describe how we do Central Logging / Monitoring at VSHN. It contains VSHN specific information, applicable to technical VSHNeers.

Central Logging

Our central logging solution is backed by Loki, which runs on APPUiO. Individual clusters ship selected logs into the central Loki database, where they can be queried through Grafana.

Each product should create their own organization in Loki to store logs they want to use for their dashboards.

Please update the list below if you create an organization for your product.

  • VSHN Managed OpenShift uses the audit organization in Loki to store Kubernetes audit logs.

  • VSHN’s Keycloak Identity Provider (id.vshn.net) uses the keycloak organization in Loki to store Keycloak audit logs.

The configuration as to which logs are shipped, and whether they should be pre-filtered, should be managed in the global Commodore defaults in the product’s distribution or other appropriate top-level class.

In general, we should be conservative with the amount of logs shipped to Loki. Logs should be shipped on an as-needed basis, and should be pre-filtered to the extent feasible.

Central Monitoring

Our central monitoring solution is backed by Mimir, which runs on APPUiO. Individual clusters ship selected metrics into the central Mimir timeseries database, where they can be queried through Grafana.

Each product should create their own organization in Mimir to store metrics they want to use for their dashboards.

Please update the list below if you create an organization for your product.

  • VSHN Managed OpenShift uses the appuio-managed-openshift-metrics organization in Mimir to store cluster monitoring metrics.

The configuration as to which metrics are shipped, and whether they should be pre-aggregated, should be managed in the global Commodore defaults in the product’s distribution or other appropriate top-level class.

In general, we should be conservative with the amount of timeseries shipped to Mimir. Timeseries should be shipped on an as-needed basis, and should be pre-aggregated to the extent feasible.