Bunny Database

Original Article ↗ Hacker News Discussion ↗

Overall reception of Bunny & new DB

  • Many are happy long‑time Bunny CDN/storage users and are excited to see a database added to the platform.
  • Others see it as another “managed SQLite in the cloud” offering and are unsure what differentiates it from Turso or Cloudflare D1.

Comparison with Cloudflare / Turso / others

  • Several view Bunny as a direct Cloudflare competitor (CDN, edge compute, DB, video, etc.).
  • Reasons cited to choose Bunny over Cloudflare D1:
    • EU‑based company and infra, appealing for data‑sovereignty and US‑vendor avoidance.
    • Better region granularity and count; explicit regional replication choices.
    • Pricing that appears substantially cheaper once outside Cloudflare’s free tier, and no Workers lock‑in (HTTP API instead).
    • Perceived better support responsiveness from a smaller company.
  • Reasons to choose Bunny over Turso:
    • Integration with Bunny’s CDN, edge scripting (Deno‑based), and “magic containers”.
    • Bunny runs its own infra rather than relying on large US clouds; preferred by some European companies.
  • Concern: libSQL (the engine used) is seen by a few as less active than upstream SQLite and with weaker driver support (e.g., Python).

Pricing & cost control

  • Pricing model (rows read/written + per‑GB per‑region) is praised as simple and low, especially versus D1.
  • Prepay and hard budget caps are highly valued to avoid surprise five‑ or six‑figure bills.
  • A minimum monthly charge on the CDN surprised at least one tester.

Capabilities, architecture & limitations

  • DB is SQLite‑compatible via libSQL; some confusion about the exact client interface (Hrana/HTTP vs “plain” SQLite).
  • Docs indicate: one writable primary, multiple read replicas that proxy writes to the primary, data stored in object storage when compute scales to zero.
  • Questions remain about suitability for write‑heavy workloads and whether it inherits SQLite’s write‑concurrency constraints; this is not clearly answered in the thread.
  • Some see it mainly fitting edge‑read / replica / per‑tenant use cases, not as a general “do‑everything” RDBMS.

Trust, reliability & product focus

  • Multiple commenters worry Bunny is stretching itself thin: lots of new platform features, with a pattern of products reaching ~80% and staying there.
  • S3 compatibility for storage is a repeated sore point:
    • Announced for 2022, then delayed, then re‑promised for early 2024; still not generally available.
    • Some report ignored support tickets about the S3 roadmap and say this eroded trust.
    • Others note S3 support is in closed beta and Bunny has said a storage rewrite was required.
  • Log delivery delays (hours to days vs promised minutes) and lack of status‑page transparency further undermine confidence for some.
  • Despite this, many still regard the core CDN/video products as “rock solid” and are open to trying the DB, but are cautious about deep lock‑in.

Managed DB vs self‑hosting debate

  • One camp argues running Postgres/MySQL on a VPS is easy, cheap, and very reliable; they question the value of managed SQLite.
  • Others counter that:
    • High availability, multi‑region, backup/restore, failover, monitoring, and security patching add substantial operational burden.
    • Delegating all that to a managed service is worth the premium, especially when you’d otherwise need specialized staff.
  • There’s agreement that backups and security (CVE tracking, firewalling) are the hard parts for DIY setups.

Miscellaneous

  • Several people clicked expecting a literal “database of bunnies” and joked about the misleading title.
  • Some are enthusiastic about Bunny as a “no‑BS” alternative with no free tier and predictable pricing, even if they never use the DB.