Skip to main content

Store

@rotorsoft/act-root

@rotorsoft/act-root / act/src / Store

Interface: Store

Defined in: libs/act/src/types/ports.ts:360

Interface for event store implementations.

The Store interface defines the contract for persistence adapters in Act. Implementations must provide event storage, querying, and distributed processing capabilities through leasing and watermark tracking.

Act includes two built-in implementations:

  • InMemoryStore: For development and testing
  • PostgresStore: For production use with PostgreSQL

Custom stores can be implemented for other databases or event log systems.

Example​

import { store } from "@rotorsoft/act";
import { PostgresStore } from "@rotorsoft/act-pg";

// Replace the default in-memory store
store(new PostgresStore({
  host: "localhost",
  port: 5432,
  database: "myapp",
  user: "postgres",
  password: "secret"
}));

const app = act()
  .withState(Counter)
  .build();

See​

  • InMemoryStore for the default implementation
  • PostgresStore for the PostgreSQL implementation

Extends​

Properties​

ack​

ack: (leases) => Promise<Lease[]>

Defined in: libs/act/src/types/ports.ts:558

Acknowledges successful processing of leased streams.

Updates the watermark to indicate events have been processed successfully. Releases the lease so other workers can process subsequent events.

Parameters​

leases​

Lease[]

Leases to acknowledge with updated watermarks

Returns​

Promise<Lease[]>

Acknowledged leases

Example​

const leased = await store().claim(5, 5, randomUUID(), 10000);
// Process events up to ID 150
await store().ack(leased.map(l => ({ ...l, at: 150 })));

See​

claim for acquiring leases

block​

block: (leases) => Promise<BlockedLease[]>

Defined in: libs/act/src/types/ports.ts:591

Blocks streams after persistent processing failures.

Blocked streams won't be returned by claim until manually unblocked. This prevents poison messages from repeatedly failing and consuming resources.

Streams are typically blocked when:

  • Max retries reached
  • blockOnError option is true
  • Handler throws an error

Parameters​

leases​

BlockedLease[]

Leases to block with error messages

Returns​

Promise<BlockedLease[]>

Blocked leases

Example​

try {
  await processEvents(lease);
  await store().ack([lease]);
} catch (error) {
  if (lease.retry >= 3) {
    await store().block([{
      ...lease,
      error: error.message
    }]);
  }
}

See​

claim for lease management

claim​

claim: (lagging, leading, by, millis, lane?) => Promise<Lease[]>

Defined in: libs/act/src/types/ports.ts:491

Atomically discovers and leases streams for reaction processing.

Atomically discovers a stream and acquires a lease in one round-trip, eliminating the race that exists when discovery and locking are separate calls (a competing worker can grab the stream between the two).

PostgresStore uses FOR UPDATE SKIP LOCKED for zero-contention competing consumer semantics — workers never block each other, each grabbing different streams atomically. InMemoryStore fuses its poll+lease logic equivalently.

Used by Act.drain() as the primary stream acquisition method.

Parameters​

lagging​

number

Max streams from the lagging frontier (ascending watermark)

leading​

number

Max streams from the leading frontier (descending watermark)

by​

string

Unique lease holder identifier (UUID)

millis​

number

Lease duration in milliseconds

lane?​

string

Optional lane filter (ACT-1103)

Returns​

Promise<Lease[]>

Array of successfully leased streams with metadata

Example​

const leased = await store().claim(5, 5, randomUUID(), 10000);
leased.forEach(({ stream, at, lagging }) => {
  console.log(`Leased ${stream} at ${at} (lagging: ${lagging})`);
});

See​

  • subscribe for registering new streams (used by correlate)
  • ack for acknowledging completion
  • block for blocking failed streams

commit​

commit: <E>(stream, msgs, meta, expectedVersion?) => Promise<Committed<E, keyof E>[]>

Defined in: libs/act/src/types/ports.ts:421

Commits one or more events to a stream atomically.

This is the core method for persisting events. It must:

  • Assign global sequence IDs to events
  • Increment the stream version
  • Check optimistic concurrency if expectedVersion is provided
  • Store events atomically (all or nothing)
  • Attach metadata (id, stream, version, created timestamp)

Type Parameters​

E​

E extends Schemas

Event schemas

Parameters​

stream​

string

The stream ID to commit to

msgs​

Message<E, keyof E>[]

Array of messages (events) to commit

meta​

EventMeta

Event metadata (correlation, causation)

expectedVersion?​

number

Expected current version for optimistic concurrency

Returns​

Promise<Committed<E, keyof E>[]>

Array of committed events with full metadata

Throws​

If expectedVersion doesn't match current version

Example​

const events = await store().commit(
  "user-123",
  [{ name: "UserCreated", data: { email: "user@example.com" } }],
  { correlation: "req-456", causation: { action: {...} } },
  0 // Expect version 0 (new stream)
);

dispose​

dispose: Disposer

Defined in: libs/act/src/types/ports.ts:26

Inherited from​

Disposable.dispose

drop​

drop: () => Promise<void>

Defined in: libs/act/src/types/ports.ts:390

Drops all data from the store.

Dangerous operation that deletes all events and state. Use with extreme caution, primarily for testing or development environments.

Returns​

Promise<void>

Example​

// Clean up after tests
afterAll(async () => {
  await store().drop();
});

notify?​

optional notify?: (handler) => NotifyDisposer | Promise<NotifyDisposer>

Defined in: libs/act/src/types/ports.ts:999

Optional cross-process commit notifications.

When implemented, the Act orchestrator subscribes once at build time and routes notifications to wake settle() automatically — so a remote worker's commit triggers reactions on this process without waiting for the debounce/poll cycle. Subscribers also receive each notification on the "notified" lifecycle event for fan-out (SSE pushes, audit logs, dashboards).

Self-filtering contract: implementations must skip their own commits. The handler fires only for commits originating from a different process writing to the same backing store. This keeps the local fast path (do() already arms drain) free of duplicate wake-ups and gives "notified" a clean cross-process semantic.

Hint, not a contract: the orchestrator never depends on notify for correctness. If absent, dropped, or the store omits it, the existing debounce/poll path still drains correctly — notify only lowers cross-process p99 reaction latency.

Adapter status (Act 0.x):

  • PostgresStore: implemented via LISTEN/NOTIFY on the act_commit channel
  • InMemoryStore: not implemented (single-process — no remote writers exist)
  • SqliteStore: not implemented (single-node by design)

Parameters​

handler​

(notification) => void

Callback invoked once per remote commit

Returns​

NotifyDisposer | Promise<NotifyDisposer>

Disposer releasing the underlying listener

prioritize​

prioritize: (filter, priority) => Promise<number>

Defined in: libs/act/src/types/ports.ts:723

Bulk-update the scheduling priority of streams matching a filter.

Used by Act.prioritize for operator runtime control over lagging-frontier claim() ordering. Unlike subscribe, which keeps the per-stream priority at the max() of all registered reactions targeting that stream, prioritize sets the priority directly to priority for matching rows — letting operators override the build-time scheduling policy.

Filter semantics mirror query_streams: stream/source are regex by default, exact with the *_exact flags. blocked restricts to blocked or unblocked rows. Omitted fields don't filter. An empty filter ({}) updates every registered stream — useful for "reset all priorities to N" but a footgun otherwise.

Parameters​

filter​

StreamFilter

PrioritizeFilter selecting which streams to update. Required (use {} to target all).

priority​

number

New priority value. Set as-is — no max(), no clamp.

Returns​

Promise<number>

Count of streams whose priority was changed.

Examples​

await store().prioritize(
  { stream: "^projection-orders$", stream_exact: false },
  10
);
await store().prioritize({ source: "^audit-" }, -5);

See​

  • Act.prioritize for the orchestrator-level wrapper
  • claim for how priority biases stream scheduling

query​

query: <E>(callback, query?) => Promise<number>

Defined in: libs/act/src/types/ports.ts:454

Queries events from the store with optional filtering.

Calls the callback for each matching event. The callback approach allows processing large result sets without loading everything into memory.

Type Parameters​

E​

E extends Schemas

Event schemas

Parameters​

callback​

(event) => void

Function invoked for each matching event

query?​

Readonly<{ after?: number; backward?: boolean; before?: number; correlation?: string; created_after?: Date; created_before?: Date; limit?: number; names?: string[]; stream?: string; stream_exact?: boolean; with_snaps?: boolean; }>

Optional filter criteria — see Query for fields (stream, name, after, before, created_after, created_before, limit, with_snaps, stream_exact).

Returns​

Promise<number>

Total number of events processed

Example​

let count = 0;
await store().query(
  (event) => {
    console.log(event.name, event.data);
    count++;
  },
  { stream: "user-123" }
);
console.log(`Found ${count} events`);

query_stats​

query_stats: <E>(input, options?) => Promise<Map<string, StreamStats<E>>>

Defined in: libs/act/src/types/ports.ts:963

Per-stream aggregated stats — single round trip per adapter.

Returns the latest event (head) plus opt-in extras (tail, count, names) for each stream selected by input. Streams with no qualifying events are absent from the result map.

Cost model. With no opt-in flags, the call uses an index-backed head lookup per stream — O(K) where K is the number of matched streams. tail alone stays in the cheap tier. Setting count and/or names triggers a full event scan over the matched streams (O(N) where N is total events); both stats share that scan, so requesting one or both is the same cost.

input. Either an explicit string[] of stream names, or a narrow event-stream selector { stream?, stream_exact? } for pattern-based or exact-name matching. Subscription-level filters (source, blocked) are intentionally not accepted here — they describe subscriptions, not events, and conflating the two would silently exclude unsubscribed event streams. For "stats for all blocked subscriptions" compose explicitly: query_streams({blocked: true}) → collect names → query_stats(names).

head vs tail naming. Follows the git-log convention: head is the latest event (highest id), tail is the earliest (lowest id). This is the opposite of the Unix head/tail commands.

Framework markers. Snapshots (__snapshot__) and tombstones (__tombstone__) are real events and are included by default — intentional, so schema-evolution tooling can count them. To exclude them, pass them in options.exclude (typed against EventName) so typos are compile-time errors.

Snapshot counts come from names. When names: true and snapshots are not in exclude, result.names["__snapshot__"] is the snapshot count for that stream — no separate field needed. Validates snapshot policy at scale: names["__snapshot__"] / count should match the configured snap predicate's expected ratio.

Time travel. options.before narrows to events with id < before, answering "what did this stream look like at event N?" without special call shape.

Type Parameters​

E​

E extends Schemas

Event schemas. Narrow at the call site to type-check exclude against your event names (typos fail at compile time).

Parameters​

input​

string[] | Pick<StreamFilter, "stream" | "stream_exact">

Stream names or a filter selecting the streams to stat.

options?​

QueryStatsOptions<E>

Opt-in stat fields, event-name exclusions, and time-travel cutoff. See QueryStatsOptions.

Returns​

Promise<Map<string, StreamStats<E>>>

Map keyed by stream name. Streams with no qualifying events (after exclude and before are applied) are absent.

Examples​

const stats = await store().query_stats(streams, {
  exclude: [TOMBSTONE_EVENT],
});
for (const [stream, { head }] of stats) {
  // head.id, head.version, head.name
}
const stats = await store().query_stats<MyEvents>(
  { stream: "^orders-" },
  { count: true, tail: true, names: true,
    exclude: [TOMBSTONE_EVENT] }
);
for (const [stream, s] of stats) {
  const snaps = s.names?.[SNAP_EVENT] ?? 0;
  const domain = (s.count ?? 0) - snaps;
  console.log(stream, { snaps, domain, tail: s.tail?.created });
}
const stats = await store().query_stats<TicketEvents>(
  { stream: "^ticket-" },
  { names: true }
);
for (const [stream, { names = {} }] of stats) {
  if ((names["TicketOpened"] ?? 0) > 0) {
    console.log(`${stream}: ${names["TicketOpened"]} legacy events`);
  }
}
const stats = await store().query_stats(["order-42"], {
  before: 100_000, // events up to (not including) id 100000
  tail: true,
});
const { head, tail } = stats.get("order-42") ?? {};
// head = latest event with id < 100_000; tail = earliest in range

See​

query_streams​

query_streams: (callback, query?) => Promise<QueryStreamsResult>

Defined in: libs/act/src/types/ports.ts:855

Streams registered subscription positions to a callback, plus the highest event id in the store.

Read-only introspection for operational dashboards (Store / Subscriptions tab, projection lag, blocked subscriptions). Avoids forcing apps to open a second connection and run raw SQL against adapter-specific schemas.

Mirrors the Store.query callback pattern: the callback is invoked once per matching position, allowing large result sets to be processed without buffering. Results are ordered by stream name; use query.after (last seen stream name) for keyset pagination on big tables (dynamic reactions can produce one subscription per aggregate).

Parameters​

callback​

(position) => void

Invoked once per matching StreamPosition.

query?​

QueryStreams

Optional QueryStreams filter (default limit: 100).

Returns​

Promise<QueryStreamsResult>

maxEventId and the count of positions emitted.

Examples​

const { maxEventId } = await store().query_streams(
  (s) => console.log(`${s.stream}: lag=${maxEventId - s.at} ${s.error}`),
  { blocked: true, limit: 50 }
);
let after: string | undefined;
for (;;) {
  const page: StreamPosition[] = [];
  const { count } = await store().query_streams(
    (s) => page.push(s),
    { after, limit: 100 }
  );
  if (!count) break;
  // ... use page ...
  after = page.at(-1)?.stream;
}

reset​

reset: (input) => Promise<number>

Defined in: libs/act/src/types/ports.ts:628

Resets watermarks for the given streams to -1, making them eligible for replay from the beginning. Also clears retry, blocked, error, and lease state so the streams can be claimed immediately.

Prefer Act.reset() over calling this directly. This primitive only resets the store; it does not raise the orchestrator's internal "needs drain" flag, so a settled Act instance will short-circuit and skip the replay. Act.reset() wraps this and arms the flag.

Accepts either an explicit list of stream names or a StreamFilter for bulk operations (e.g., "rebuild every blocked stream"). The filter form is the same shape used by prioritize and query_streams. An empty filter ({}) matches every registered stream — typically a footgun for reset; prefer narrower filters like { blocked: true }.

Parameters​

input​

string[] | StreamFilter

Stream names or a StreamFilter

Returns​

Promise<number>

Count of streams that were actually reset

Example​

// By name
await app.reset(["my-projection"]);

// By filter — rebuild every blocked stream in a projection family
await app.reset({ stream: "^proj-orders-", blocked: true });

// Low-level (does NOT trigger replay on settled apps)
await store().reset(["my-projection"]);

See​

Act.reset for the high-level rebuild API that wraps this primitive and arms the orchestrator's drain flag

restore?​

optional restore?: (driver) => Promise<void>

Defined in: libs/act/src/types/ports.ts:807

Atomically wipe the store and commit a fresh sequence of events inside a single transaction.

Capability-gated. Adapters may or may not implement restore. Third-party stores that can't atomically wipe-and-rebuild in one transaction can omit it.

Driver pattern. The adapter is a thin transactional wrapper: open the transaction (PG BEGIN, SQLite BEGIN IMMEDIATE, an in-process snapshot for InMemoryStore), truncate the events + streams/subscriptions tables, hand the orchestrator a per-event insert callback by invoking driver(callback), then commit or roll back. Any throw inside driver rolls back the transaction — the store ends byte-for-byte unchanged from the pre-call state.

The framework's scan loop (in internal/event-sourcing.ts) is what calls callback repeatedly: it iterates the source, validates each event, applies drop_snapshots, fires on_progress, rewrites meta.causation.event.id through the per-call old → new map, and counts kept/dropped. Adapters never see that logic — their job is the transaction lifecycle plus the adapter-specific callback body.

Lossless created. The callback receives the event's original timestamp; adapters write it through verbatim. This is the property that makes restore a viable backup/migration primitive — distinct from commit, which always stamps now().

Renumbered id. Adapters reseed ids densely (1..N on PG/SQLite, 0..N-1 on InMemory). The source's original ids are used by the orchestrator as causation lookup keys but never written through.

No subscription preservation. Both the events and the streams/subscriptions tables are wiped. Reactions re-subscribe via the orchestrator's normal correlate() path on the next settle cycle.

Cache. Restore does not touch the Cache port — callers must cache().clear() after restore to avoid serving stale snapshots. Documented; not enforced.

Parameters​

driver​

(callback) => Promise<void>

Orchestrator-supplied iteration callback. The adapter calls driver(callback) exactly once, from inside its transaction. The callback argument is the adapter's per-event insert hook — it receives the event with meta.causation already rewritten to the new id space and returns the new id the adapter assigned. The driver is purely transactional from the adapter's perspective — kept/dropped counts and timing live in Act.restore.

Returns​

Promise<void>

See​

  • Act.restore for the public entry point.
  • truncate for the single-stream snapshot/tombstone primitive (different operation — restore wipes the whole store)

seed​

seed: () => Promise<void>

Defined in: libs/act/src/types/ports.ts:375

Initializes or resets the store.

Used primarily for testing to ensure a clean state between tests. For production stores, this might create necessary tables or indexes.

Returns​

Promise<void>

Example​

// Reset store between tests
beforeEach(async () => {
  await store().seed();
});

subscribe​

subscribe: (streams) => Promise<{ subscribed: number; watermark: number; }>

Defined in: libs/act/src/types/ports.ts:522

Registers streams for event processing.

Upserts stream entries so they become visible to claim. Used by correlate() to register dynamically discovered reaction target streams.

Also returns the current maximum watermark across all subscribed streams, used internally for correlation checkpoint initialization on cold start.

Parameters​

streams​

object[]

Streams to register with optional source hint

Returns​

Promise<{ subscribed: number; watermark: number; }>

subscribed count of newly registered streams, watermark max at across all streams

Example​

const { subscribed, watermark } = await store().subscribe([
  { stream: "stats-user-1", source: "user-1" },
  { stream: "stats-user-2", source: "user-2", priority: 10 },
]);

See​

  • claim for discovering and leasing registered streams
  • prioritize for changing priority after subscription

truncate​

truncate: (targets) => Promise<TruncateResult>

Defined in: libs/act/src/types/ports.ts:741

Atomically truncates streams and seeds each with a final event.

For each target, in a single transaction:

  1. Deletes all events for the stream
  2. Removes the stream's entry from the streams table
  3. Inserts a __snapshot__ (when snapshot is provided) or __tombstone__ event as the sole event on the stream

Parameters​

targets​

object[]

Streams to truncate with optional snapshot state and meta

Returns​

Promise<TruncateResult>

Map keyed by stream name, each entry with deleted count and committed event

See​

Act.close for the high-level close-the-books API that orchestrates safety checks, archive callbacks, and atomic truncate+seed

unblock​

unblock: (input) => Promise<number>

Defined in: libs/act/src/types/ports.ts:682

Clears the blocked flag on streams without replaying their history. Sets blocked = false, retry_count = 0, error = null, and clears any lease bookkeeping. The at watermark stays where it was — the stream resumes from the next event after the last successful ack, not from zero.

The distinction from reset matters: reset() is for projection rebuilds (replay from event 0); unblock() is for recovering from a poison message after the operator fixes the underlying issue. Use unblock() when you don't want to re-process history.

Prefer Act.unblock() over calling this directly. Like reset(), this primitive doesn't raise the orchestrator's internal "needs drain" flag — a settled Act instance will short-circuit and skip the resume. Act.unblock() wraps this and arms the flag.

Only streams that were actually blocked at call time count toward the return value; already-unblocked streams and unknown stream names are silently skipped. The atomic single-statement update makes the call safe to issue concurrently with claim() — workers holding a FOR UPDATE SKIP LOCKED lock won't see partial state.

Accepts either an explicit list of stream names or a StreamFilter for bulk recovery (e.g., "unblock every blocked order projection"). The blocked = true predicate is always applied — passing blocked: false in the filter matches nothing. An empty filter ({}) means "unblock everything that's blocked," which is a sane post-incident bulk recovery.

Parameters​

input​

string[] | StreamFilter

Stream names or a StreamFilter

Returns​

Promise<number>

Count of streams that were actually flipped (were blocked)

Example​

// By name (single targeted recovery)
await app.unblock(["webhooks-out-customer-42"]);

// By filter — unblock every blocked stream in a family
await app.unblock({ stream: "^webhooks-out-" });

// Post-incident: unblock everything that's blocked
await app.unblock({});

// Low-level (does NOT trigger resume on settled apps)
await store().unblock(["webhooks-out-customer-42"]);

See​

  • Act.unblock for the high-level recovery API
  • reset for the rebuild-from-zero alternative