Where sites and services live. Pick the host for the workload, then layer Cloudflare and DNS on top.
Hosting
- github-pages: Static hosting from Actions; limits, custom domain, HTTPS.
- vercel: Setup, Node.js on Fluid compute, previews, ISR, domains.
- vercel-env-vars: Env scoping, Config vs Secret types, size limits, rotation.
- vercel-preview-deployments: Preview URLs, Deployment Protection defaults and methods.
- vercel-edge-functions: Edge runtime status, limits, and migrating to Node.js.
- vercel-analytics: Web Analytics and Speed Insights setup and event pricing.
- hostinger-vps: Deploy by tag, systemd, journalctl, and Hostinger’s own backups.
- hostinger-vps-hardening: Key-only SSH checked with
sshd -T, UFW, fail2ban, upgrades. - hostinger-vps-reverse-proxy-caddy: Caddy install, Caddyfile, and certificates behind Cloudflare.
- hostinger-vps-backups: restic off-box backups, retention, integrity checks, restore drills.
Cloudflare and DNS
- cloudflare: Proxying, SSL/TLS Full (strict), Rules, Workers over Pages.
- cloudflare-dns: Proxied vs DNS-only, CNAME flattening, TTL auto, DNSSEC.
- cloudflare-cache-rules: Defaults, rule quotas, tiered cache, cf-cache-status, purging.
- cloudflare-security-headers: HSTS, CSP, and friends via Response Header Transform Rules.
- cloudflare-workers: Limits and pricing, fetch handler, bindings, wrangler, Node compat.
- cloudflare-kv: Eventually consistent KV, write limits, bulk operations.
- cloudflare-r2: S3-compatible storage, prices, multipart, lifecycle, presigned URLs.
- cloudflare-durable-objects: SQLite-backed actors, hibernating WebSockets, alarms.
- namecheap-dns: Record types, ALIAS at the apex, MX, TTL, delegating to Cloudflare.
Delivery and data
- ci-cd: Gate order, GitHub Actions, caching, required checks, hooks.
- postgres-prod: Managed vs self-hosted, restore drills, PITR, failover, pooling.
- supabase-backups: Daily backups by plan, PITR trade-offs, nightly pg_dump offsite.
- secrets-and-env: Client vs server variables, Supabase key types, scanning, rotation.
Reliability and launch
- error-tracking: Sentry init, source maps with debug IDs, releases, PII scrubbing, alerts.
- incident-response: Severities, escalation, roles, and the first ten minutes.
- blameless-postmortems: When to write one, the five-part structure, template.
- disaster-recovery: RPO and RTO, offsite copies, dead-man’s switch, restore drills.
- pre-launch-checklist: Day-one breakage, six security checks, legal, error tracking.
LLM operations
- llmops-best-practices: Pillar: versioned prompts, eval gates, tracing, cost, safety, rollback.
- llm-evaluation-in-production: Golden sets, CI gates, judge calibration, model-change re-evals.
- llm-observability: Per-request traces, token and cost metrics, quality signals, alerts.