https://k-s-pavan-kumar.github.io/hugo-nimbus/adr/Recent content in Architecture Decision Records on NimbusHugoen-usSat, 01 Jun 2024 00:00:00 +0000https://k-s-pavan-kumar.github.io/hugo-nimbus/adr/0002-gitops-cluster-config/Sat, 01 Jun 2024 00:00:00 +0000https://k-s-pavan-kumar.github.io/hugo-nimbus/adr/0002-gitops-cluster-config/<h2 id="context">Context</h2> <p>Cluster configuration changes currently happen via direct <code>kubectl apply</code> from engineer laptops. We have no audit trail and drift between clusters is common.</p> <h2 id="decision">Decision</h2> <p>Adopt a GitOps workflow (Argo CD) where the Git repository is the single source of truth for cluster state, and all changes go through pull request review.</p> <h2 id="consequences">Consequences</h2> <ul> <li>Every config change is reviewable and revertible via Git history.</li> <li>Emergency changes need a documented break-glass procedure, since direct <code>kubectl apply</code> will be restricted in production namespaces.</li> <li>Requires migrating ~40 existing manifests into the GitOps repo before enforcement can begin.</li> </ul>https://k-s-pavan-kumar.github.io/hugo-nimbus/adr/0001-postgresql-primary-datastore/Sat, 10 Feb 2024 00:00:00 +0000https://k-s-pavan-kumar.github.io/hugo-nimbus/adr/0001-postgresql-primary-datastore/<h2 id="context">Context</h2> <p>We need a primary datastore for the new billing service. The team has production experience with both PostgreSQL and MongoDB, and the data model is relational (accounts, invoices, line items with foreign keys).</p> <h2 id="decision">Decision</h2> <p>We will use PostgreSQL 16, managed via RDS, as the primary datastore for all new services unless a specific workload has a strong document-store or time-series justification.</p> <h2 id="consequences">Consequences</h2> <ul> <li>Schema migrations require a formal review (see <code>docs/migrations</code>).</li> <li>We standardize on one connection-pooling strategy (PgBouncer) across services, simplifying on-call runbooks.</li> <li>Document-shaped data (e.g. audit logs) will need a deliberate exception request rather than defaulting to Postgres JSONB columns.</li> </ul>