Hyperdrive (Postgres)

in-progress

Hyperdrive gives Workers a low-latency connection pool to your Postgres. NextDeploy declares the config, wires up the binding, and injects the connection string at deploy time. Currently a Tier A blocker — see status.

Source reference: cli/internal/serverless/cloudflare_hyperdrive.go, CLOUDFLARE_PARITY.md

What Hyperdrive solves

TODO Workers run anywhere — opening Postgres connections cold from each ISO is slow. Hyperdrive is a managed proxy / pool that keeps warm connections globally.

Status today

TODO binding declaration works; auto-injection of the connection string into the worker is a Tier A blocker — for now you supply it manually as a secret and reference it in code.

Declaring in nextdeploy.yml

TODO YAML shape — name, origin (your Postgres URL via secret), caching, max connections.

Connection string as a secret

TODO put your Postgres URL in nextdeploy secrets set DB_URL=.... Hyperdrive resolves it server-side; your code only sees the binding.

Caching reads

TODO Hyperdrive cache layer for read-heavy workloads, TTL config, when NOT to cache (mutating tables).

Using from your app

TODO code example with pg (or compatible driver) talking to env.HYPERDRIVE.connectionString. Latency expectations.

Limitations

TODO TCP only (no TLS termination yet in some cases), max connection limits, no DDL via Hyperdrive (apply migrations directly).

Pricing

TODO per-query pricing, free tier numbers.