Bahriya Tersane

The catalogue

Every chart in Tersane.

The charts that run the managed services on bahriya.cloud, grouped by what they provide. Every entry installs with Helm; where a service is best run by a Kubernetes operator, its chart installs the operator. Software with an upstream name keeps it, so you know what you are running; the charts Mamluk designed for deploying your own applications carry Arabic names.

Data

  • mysql MySQL, standalone or as a highly available clusterDocumentation
  • postgresql PostgreSQLDocumentation
  • cloudnative-pg CloudNativePG, the operator that runs the PostgreSQL clusters, installed once per clusterDocumentation
  • valkey Valkey, standalone, with Sentinel, or shardedDocumentation
  • memcached MemcachedDocumentation
  • couchdb Apache CouchDB, a document databaseDocumentation
  • opensearch OpenSearch, for search and log analyticsDocumentation
  • velero Velero, for backing up and restoring cluster resources and volumesUnder consideration

Messaging

Identity

  • keycloak Keycloak, for identity and single sign-onDocumentation

Registry and CI

  • harbor Harbor, a container image and chart registryDocumentation
  • onedev OneDev, for Git repositories with code review, issues, and container, npm and Maven registriesDocumentation
  • concourse Concourse, for continuous integration and deliveryDocumentation

Workloads

  • safeenaship Web services, with ingress, TLS and autoscalingDocumentation
  • khadimservant Workers and daemonsDocumentation
  • saathour Scheduled jobsDocumentation
  • sipar-gatewayshield The gateway for a cluster, installed once per cluster, with address-based access control for TCP clients and a response cache, each behind a switchDocumentation
  • sipar-routeshield A library chart, used inside an application's own chart and never installed on its own, that renders the application's entry on the cluster's Sipar gateway, with its access, rate-limit and cache policies; it needs sipar-gatewayDocumentation

Security

  • clamav ClamAV, as a shared malware scanner in the clusterDocumentation
  • openbao OpenBao, for secrets managementDocumentation
  • secrets-store-csi-driver The Secrets Store CSI driver, which mounts secrets into Pods as files through a providerDocumentation
  • kyverno Kyverno, for admission policyDocumentation
  • policy-reporter Policy Reporter, a view in the cluster of Kyverno's policy reportsDocumentation
  • trivy-operator Trivy Operator, for scanning the images running in the clusterDocumentation
  • logging Cluster logs in OpenSearch, with the Kubernetes audit log, dashboards and alertsDocumentation

How the charts are made

Defaults decided for production.

Each chart carries the values bahriya.cloud runs, and each default is a decision you can read and override in a values file you keep. Among them:

  • Requests and limits on every container. In the data and workload charts, init containers, sidecars and test pods included, so a release does not fail against a namespace quota.
  • Spread across nodes. Replicas are spread across nodes wherever the cluster allows, so the loss of one node does not take a service with it.
  • Disruption budgets. A PodDisruptionBudget by default for long-running workloads and caches, so a node drain cannot remove every replica at once.
  • Network policy for the data charts. MySQL and Valkey are reachable from their own namespace by default, and from elsewhere only where your values say so.
  • Backups and recovery built in. MySQL and Valkey back up to S3-compatible storage when backups are enabled, and their documentation sets out the restore, including point-in-time recovery for MySQL.

Each chart's documentation sets out its defaults in full.

Industry-specific additions, such as the Banking Convoy, are described under Convoys.