The best way to control Claude Code, Codex and more from any computer or phone
Mirrored from https://github.com/Camo67/t3code.git
Synced Nov 21, 06:00 UTC
| 1 | --- |
| 2 | title: State Store |
| 3 | description: How Alchemy persists resource state between deploys to compute diffs and track infrastructure. |
| 4 | --- |
| 5 | |
| 6 | import StateStoreBootstrap from "../../../components/marketing-islands/StateStoreBootstrap.tsx"; |
| 7 | |
| 8 | Alchemy persists resource state between deploys so it can compute |
| 9 | diffs — comparing the desired state in your code against the current |
| 10 | state of your infrastructure. |
| 11 | |
| 12 | ## How state is stored |
| 13 | |
| 14 | Each resource's state is keyed by its **fully qualified name** (FQN), |
| 15 | which includes the namespace path and |
| 16 | [logical ID](/infrastructure-as-code/resource#logical-id). State is |
| 17 | scoped by [stack name](/infrastructure-as-code/stack) and |
| 18 | [**stage**](/environments/stages), so different stacks and |
| 19 | environments are fully isolated. |
| 20 | |
| 21 | A resource's persisted state includes: |
| 22 | |
| 23 | - **Resource type** — e.g. `Cloudflare.R2.Bucket` |
| 24 | - **Input properties** — the props you passed when creating it |
| 25 | - **Output attributes** — the values returned after creation |
| 26 | - **Instance ID** — a unique identifier for this instance |
| 27 | - **Lifecycle status** — `created`, `updating`, `deleting`, etc. |
| 28 | - **[Bindings](/infrastructure-as-effects/binding)** — data attached by policies and event sources |
| 29 | |
| 30 | ## Local state |
| 31 | |
| 32 | By default, state is stored on disk in the `.alchemy/` directory: |
| 33 | |
| 34 | ``` |
| 35 | .alchemy/ |
| 36 | state/ |
| 37 | MyApp/ |
| 38 | dev_sam/ |
| 39 | Bucket.json |
| 40 | Worker.json |
| 41 | ``` |
| 42 | |
| 43 | Add `.alchemy/` to your `.gitignore`: |
| 44 | |
| 45 | ```sh |
| 46 | echo ".alchemy/" >> .gitignore |
| 47 | ``` |
| 48 | |
| 49 | Local state works for solo development. Each developer gets their own |
| 50 | state via their default stage (`dev_$USER`). |
| 51 | |
| 52 | ## Remote state |
| 53 | |
| 54 | For teams and CI, configure a remote state store so all deploys share |
| 55 | the same state. Alchemy includes a Cloudflare-backed store: |
| 56 | |
| 57 | ```typescript |
| 58 | Alchemy.Stack( |
| 59 | "MyApp", |
| 60 | { |
| 61 | providers: Cloudflare.providers(), |
| 62 | state: Cloudflare.state(), |
| 63 | }, |
| 64 | Effect.gen(function* () { |
| 65 | // ... |
| 66 | }), |
| 67 | ); |
| 68 | ``` |
| 69 | |
| 70 | ## Cloudflare state store |
| 71 | |
| 72 | `Cloudflare.state()` is the recommended state store for projects |
| 73 | already running on Cloudflare. It persists state in a Worker backed |
| 74 | by a Durable Object with embedded SQLite, with the auth token and |
| 75 | encryption key kept in your account's Secrets Store. |
| 76 | |
| 77 | ### First-run bootstrap |
| 78 | |
| 79 | The first time you run `alchemy deploy`, `plan`, or `dev` against a |
| 80 | stack configured with `Cloudflare.state()`, Alchemy can't find the |
| 81 | state-store Worker on your account and deploys it, once, with its |
| 82 | supporting resources: |
| 83 | |
| 84 | <StateStoreBootstrap client:visible /> |
| 85 | |
| 86 | What gets created: |
| 87 | |
| 88 | - **`Api`** — the state-store Worker itself (a Durable Object with |
| 89 | embedded SQLite, enabled on a `workers.dev` subdomain) |
| 90 | - **`AlchemyStateStoreToken`** — bearer token used by the CLI to |
| 91 | authenticate to the Worker, stored in your Secrets Store |
| 92 | - **`StateStoreEncryptionKey`** — symmetric key used to encrypt |
| 93 | resource state at rest inside the Durable Object's SQLite |
| 94 | - **`StateStoreSecrets`** — the Secrets Store that holds the two |
| 95 | secrets above, scoped to your Cloudflare account |
| 96 | - **`StateStoreAuthTokenValue` / `StateStoreEncryptionKeyValue`** — |
| 97 | the generated random values bound into the secrets above |
| 98 | |
| 99 | These resources are reused across every stack and stage that uses |
| 100 | `Cloudflare.state()` on this account — no duplication per project. |
| 101 | Subsequent runs detect the existing Worker and use it. |
| 102 | |
| 103 | In CI (`CI=true`) a missing store fails the run instead: a deploy there |
| 104 | should find its store provisioned, not create one as a side effect. |
| 105 | Provision it once ahead of time. Without `--backend`, the command |
| 106 | provisions the store your `alchemy.run.ts` configures: |
| 107 | |
| 108 | ```sh |
| 109 | alchemy state bootstrap --profile <your-ci-profile> |
| 110 | ``` |
| 111 | |
| 112 | `AWS.state()` behaves the same way: it creates its DynamoDB table on |
| 113 | first use, except in CI, where `alchemy state bootstrap --backend aws` |
| 114 | provisions it. See [AWS state store](#aws-state-store). |
| 115 | |
| 116 | `GCP.state()` does the same with a Cloud Storage bucket in your project, |
| 117 | `alchemy-state-<project>`. In CI, provision it with |
| 118 | `alchemy state bootstrap --backend gcp`. |
| 119 | |
| 120 | ### Where credentials live |
| 121 | |
| 122 | After the bootstrap, Alchemy writes the state-store URL and bearer |
| 123 | token to a credentials file under your Alchemy profile directory |
| 124 | (`~/.alchemy/profiles/<profile>/state-stores/cloudflare.json`). On |
| 125 | CI, set `CI=true`; the credentials are resolved from the Secrets |
| 126 | Store on every run via a short-lived edge-preview Worker, so nothing |
| 127 | needs to be persisted to disk. |
| 128 | |
| 129 | ### More than one store per account |
| 130 | |
| 131 | By default the state-store Worker is named `alchemy-state-store`. Pass |
| 132 | `workerName` for a separate store, for example one per team. It gets its own |
| 133 | Worker, its own secrets in the Secrets Store, and its own cached credentials: |
| 134 | |
| 135 | ```typescript |
| 136 | state: Cloudflare.state({ workerName: "team-a-state" }), |
| 137 | ``` |
| 138 | |
| 139 | Address it from the CLI with `--name`, which takes the Worker name: |
| 140 | |
| 141 | ```sh |
| 142 | alchemy state list --backend cloudflare --name team-a-state |
| 143 | ``` |
| 144 | |
| 145 | `AWS.state({ name })` does the same with a separate DynamoDB table, |
| 146 | `alchemy-state-<name>`; pass `tableName` to choose the full table name |
| 147 | yourself. `GCP.state({ name })` uses `alchemy-state-<name>-<project>`. |
| 148 | |
| 149 | ### Pin the account |
| 150 | |
| 151 | The store lives in the Cloudflare account your profile resolves. If your |
| 152 | credentials reach several accounts, pin it, so logging in to a different one |
| 153 | fails instead of reading or deploying the store there: |
| 154 | |
| 155 | ```typescript |
| 156 | state: Cloudflare.state({ accountId: "c8eeff0e..." }), |
| 157 | ``` |
| 158 | |
| 159 | ### Serve it on your own domain |
| 160 | |
| 161 | The store answers on its `workers.dev` URL. Some networks can't reach |
| 162 | `workers.dev`, so give it a hostname in one of your zones instead. Alchemy |
| 163 | attaches it to the store's Worker as a custom domain and uses it from then on: |
| 164 | |
| 165 | ```typescript |
| 166 | state: Cloudflare.state({ domain: "alchemy-state.example.com" }), |
| 167 | ``` |
| 168 | |
| 169 | ## AWS state store |
| 170 | |
| 171 | `AWS.state()` keeps state in one on-demand DynamoDB table per account and region, `alchemy-state`. A deploy loads a stage's records with one query instead of reading an object per resource. |
| 172 | |
| 173 | ```typescript |
| 174 | Alchemy.Stack( |
| 175 | "MyApp", |
| 176 | { |
| 177 | providers: AWS.providers(), |
| 178 | state: AWS.state(), |
| 179 | }, |
| 180 | Effect.gen(function* () { |
| 181 | // ... |
| 182 | }), |
| 183 | ); |
| 184 | ``` |
| 185 | |
| 186 | A DynamoDB item holds at most 400 KB. A record larger than `overflowThreshold` (300 KiB by default) is written to an S3 overflow bucket, `alchemy-state-overflow-<account>-<region>-an`, and its item keeps the object's key. |
| 187 | |
| 188 | The store creates its table, with deletion protection, and the overflow bucket on first use, except in CI. Provision them ahead of time with: |
| 189 | |
| 190 | ```sh |
| 191 | alchemy state bootstrap --backend aws --profile <your-ci-profile> |
| 192 | ``` |
| 193 | |
| 194 | `AWS.state({ name: "team-a" })` uses a separate table, `alchemy-state-team-a`. `kmsKeyId` encrypts the table and the overflow bucket with your own KMS key. |
| 195 | |
| 196 | The credentials that deploy need DynamoDB access to the table and S3 access to the overflow bucket. The first deploy of a stage also lists the old S3 store's bucket, to check the stage isn't still there; `migrateFrom: false` skips that. |
| 197 | |
| 198 | ### Migrating from the S3 store |
| 199 | |
| 200 | Earlier versions kept AWS state in an S3 bucket, `alchemy-state-<account>-<region>-an`. A stage still there fails to deploy until you move it. Run this in the project, so it migrates the store your entrypoint configures: |
| 201 | |
| 202 | ```sh |
| 203 | alchemy state migrate --profile <your-profile> |
| 204 | ``` |
| 205 | |
| 206 | The command moves every stage, or the ones `--stack` and `--stage` select. It closes each moved stage's S3 objects in the bucket policy, so every older alchemy, say in a CI job you haven't upgraded, fails at its first read instead of deploying from the old copy. The S3 objects stay where they are. A stage you create in DynamoDB afterwards is locked in S3 too, so a teammate still on the S3 store can't create it a second time. |
| 207 | |
| 208 | A stage another deploy is running is skipped and reported. Run the command again once that deploy finishes; it picks up where it stopped, including after a crash. |
| 209 | |
| 210 | To go back to an older alchemy, move the stages back first: |
| 211 | |
| 212 | ```sh |
| 213 | alchemy state migrate --rollback |
| 214 | ``` |
| 215 | |
| 216 | Roll back with the version you migrated with, before downgrading. State written by a newer version may use fields an older one can't read. |
| 217 | |
| 218 | ### Keeping state in S3 |
| 219 | |
| 220 | `AWS.state({ backend: "s3" })` keeps using the S3 bucket. So does any config that passes `bucketName`, `prefix` or `encryption`, which only the S3 store takes. Bootstrap it with `--backend aws-s3`. |
| 221 | |
| 222 | ## Postgres state store |
| 223 | |
| 224 | `postgresState()` persists state in a Postgres database you already run, and it locks: each `(stack, stage)` holds a session advisory lock for the duration of a run, so two concurrent deploys of the same stage cannot interleave. |
| 225 | |
| 226 | ```typescript |
| 227 | import { postgresState } from "@alchemy.run/sql/PostgresState"; |
| 228 | import * as Config from "effect/Config"; |
| 229 | |
| 230 | Alchemy.Stack( |
| 231 | "MyApp", |
| 232 | { |
| 233 | providers: Prisma.providers(), |
| 234 | state: postgresState({ url: Config.Redacted("STATE_DATABASE_URL") }), |
| 235 | }, |
| 236 | Effect.gen(function* () { |
| 237 | // ... |
| 238 | }), |
| 239 | ); |
| 240 | ``` |
| 241 | |
| 242 | It is imported from `@alchemy.run/sql/PostgresState` rather than the `State` barrel because it depends on the optional `@effect/sql-pg` driver, which must stay out of bundled worker code. Install `@effect/sql-pg` to use it. |
| 243 | |
| 244 | ## State and lifecycle |
| 245 | |
| 246 | State transitions track the lifecycle of each resource: |
| 247 | |
| 248 | | Status | Meaning | |
| 249 | | ----------- | ---------------------------------------------------------------- | |
| 250 | | `creating` | Create operation in progress | |
| 251 | | `created` | Resource exists and is healthy | |
| 252 | | `updating` | Update operation in progress | |
| 253 | | `updated` | Resource was updated successfully | |
| 254 | | `deleting` | Delete operation in progress | |
| 255 | | `replacing` | Resource is being replaced (new one created, old pending delete) | |
| 256 | |
| 257 | These intermediate states (`creating`, `updating`, `deleting`) exist |
| 258 | because state is persisted before the cloud operation completes. If |
| 259 | the process crashes mid-operation, Alchemy uses the intermediate |
| 260 | state to recover on the next deploy. |
| 261 | |
| 262 | ## In-memory state (testing) |
| 263 | |
| 264 | For tests, Alchemy provides an in-memory state store so tests don't |
| 265 | touch the filesystem: |
| 266 | |
| 267 | ```typescript |
| 268 | import * as TestState from "@alchemy.run/test/TestState"; |
| 269 | |
| 270 | // Seed with existing resource state |
| 271 | const state = TestState.state({ |
| 272 | Bucket: { |
| 273 | /* ... */ |
| 274 | }, |
| 275 | }); |
| 276 | ``` |
| 277 | |
| 278 | The test harness (`@alchemy.run/test/Bun` and `@alchemy.run/test/Vitest`) |
| 279 | handles state setup automatically. |
| 280 | |
| 281 | ## Use a store without a stack file |
| 282 | |
| 283 | A provider's store is reachable from your profile alone, so scripts, CI jobs |
| 284 | and one-off tools can read and manage state with no `alchemy.run.ts`. From the |
| 285 | CLI, pass `--backend` to any `state` command: |
| 286 | |
| 287 | ```sh |
| 288 | alchemy state bootstrap --backend cloudflare --profile prod # create or upgrade it |
| 289 | alchemy state list --backend cloudflare --profile prod |
| 290 | alchemy state logs --backend cloudflare --name team-a --tail |
| 291 | ``` |
| 292 | |
| 293 | From code, use the `provider` source of [`Alchemist.State`](/cli/api#state): |
| 294 | |
| 295 | ```typescript |
| 296 | import * as Alchemist from "alchemy/Alchemist"; |
| 297 | import * as Effect from "effect/Effect"; |
| 298 | |
| 299 | const source = { type: "provider", backend: "cloudflare", profile: "prod" } as const; |
| 300 | |
| 301 | const program = Effect.gen(function* () { |
| 302 | yield* Alchemist.State.bootstrap(source); |
| 303 | const state = yield* Alchemist.State.store(source); |
| 304 | return yield* state.listStacks(); |
| 305 | }).pipe(Effect.provide(Alchemist.layer()), Effect.scoped); |
| 306 | ``` |
| 307 | |
| 308 | `Alchemist.State.layer(source)` provides the same store as a `State` layer. |
| 309 | The backend is the provider's name: `cloudflare`, `aws` or `gcp`. Pass `name` |
| 310 | for a [named store](#more-than-one-store-per-account). |
| 311 | |
| 312 | Only the profile and `name` pick the store this way. Options set in the stack |
| 313 | file, such as a [pinned account](#pin-the-account) or a |
| 314 | [custom domain](#serve-it-on-your-own-domain), aren't read; use the |
| 315 | `configured` source (the CLI default, without `--backend`) when they matter. |
| 316 | |
| 317 | ## Writing your own state store |
| 318 | |
| 319 | A state store is just an Effect `Layer` that provides the `State` |
| 320 | service — back it with Postgres, S3, Redis, DynamoDB, or any other |
| 321 | backend. See [Custom State Store](/state-store/custom-state-store) for |
| 322 | the full walkthrough. |
| 323 | |
| 324 | ## Where next |
| 325 | |
| 326 | - [Custom State Store](/state-store/custom-state-store) — implement the |
| 327 | `StateService` interface end-to-end. |
| 328 | - [Testing Providers](/testing/testing-providers) — the in-memory store in practice. |
| 329 | - [References](/infrastructure-as-code/references) — reading another |
| 330 | stack's outputs. |