Files
PaperClipAI/server/src/services/heartbeat/queue.test.ts
T
DottaandPaperclip 27d35314be feat: connect experimental Cursor Cloud to native Runner (#18069)
## Thinking Path

> - Paperclip helps people manage AI agents at work.
> - Native Runner owns execution and task authority.
> - Operators need an explicit native Cursor Cloud choice.
> - Tools/files need company, task, session and run scope.
> - This connects setup and lifecycle to Runner Core.
> - Legacy agents retain their selected runtime/history.

## Linked Issues or Issue Description

Depends on #18068. Managed-only API placement and API-owned Stop
containment are in #18310. Refs #5664, #11803, #17299, and #17471. No
duplicate native integration was found.

## What Changed

- Add explicit native setup, Test, saved settings and dispatch with
no-prompt account/catalog/callback checks.
- Scope MCP to company, agent, task, session and run through
authenticated HTTPS. Register verified artifacts/work products; preserve
unpriced accounting, completion and budgets.
- Connect Question/Plan continuation, cold handoff and admitted fresh
sessions; recheck waiting admission under the task lock.
- Present audited Stop intent while preserving physical containment
proof; bound typed Stop metadata. Preserve title edits through
preference/navigation readiness.
- Select one live/history representation per run to prevent duplicate
settled cards. Integrate upstream task/read-state/retry/schema contracts
and repair two test fixtures.

## Verification

Core #18068 merged normally as `3910352c923a`. Product head
`5a60f541c982` has passing repository CI/build/browser checks, including
required `ci / verify` and `ci / e2e`, plus a fresh exact-head Greptile
5/5 review with zero unresolved threads. Two non-required external app
checks remain pending: Paperclip Review is queued waiting for its
assigned agent; Superagent Security Scan is reviewing the diff. These
are not claimed completed. The normal merge endpoint enforces the
repository rules; no policy or status override is used. The dependent
Top PR is addressing its fresh deferred-Stop review finding.

Product `5a60f541c982` passed 270 UI and 20 selected
Cloud/native/disk-full tests, UI/server typechecks, tokens and maps.
Top’s clean normal merge passed tokens/maps/diff checks; its matrix is
pending. Earlier receipts retain Core’s 47 Rust, 4 compiled-native
retirement and 19 settlement controls, normal Runner build, and runtime
9af build/recursive typecheck. Reproduced loading/shutdown CI fixture
failures passed 7/7 browser and 121/121 unit tests, including a
broken-unsubscribe negative control. No timeout/oracle was relaxed.

The actual-parent restack adopts 75 mainline paths, including
schema/permission and generic runtime changes, beyond the prior two
fixture repairs and client projection fix. The full difference from
hosted `9af1a5de42a2` is therefore larger than four files; backend
source also differs. The merged config-read checks match Core391, and
the Cursor placement/runtime/relay/selection modules match frozen9af;
selected affected regressions pass. Paid proof keeps its actual 9af
source/image identity. The separately tested live/history card fix is
not claimed deployed on that host.

On 9af, original saved settings survived reopening and fresh no-prompt
Test authenticated catalog/model/binding; no new Save is claimed. Fresh
Linux admission passed. With `claude-sonnet-4-5`, Artifact downloaded
exact 27 bytes and matched 121 canonical identities in complete
S3-served history. First Stop’s strict quiet FAILED on a 10.373-second
gap despite independent cancellation/no late artifact. One targeted
unchanged-sampler retry retained all 24 samples: 35.076 seconds quiet /
maximum gap 4.584 seconds, 67 ordered canonical identities, and separate
CANCELLED/artifact-absent status after the original 180-second deadline.
The pre-reload duplicate settled card is disclosed. Exact-ID cleanup for
the three new provider agents is awaiting authorization, so full grades
and the fresh Artifact’s post-cleanup download remain blocked. A
historical deleted-provider artifact was newly downloaded on 9af;
original F226 Question/Plan/reconnect and cleanup proof keep their
identities.

Retained failures: stale fixture 132/133; E8 48/135 with 87 reset-hook
failures; Mac 17 SQL failures/skips; earlier
title/startup/custody/Stop/deletion failures with unknown causes/bounded
recoveries; `dea023` Stop’s warning/50 duplicate identities; pre-repair
ACK and original Core CI failures. WSP’s 251-assertion, 7/8-suite
attempt failed physical PostgreSQL teardown; the unchanged wrapper rerun
passed 128. Retry alone establishes no cause or full local-suite pass.
All paid attempts remain recorded; vendor dollar cost is UNKNOWN.

## Risks

Experimental: Cursor independently acknowledges neither the effective
model nor instruction digest before the first prompt. Explicit values
are submitted and checked through behavior; unpriced provider dollar
cost stays UNKNOWN.

Callbacks need authenticated reachable HTTPS. Scope, revocation,
budgets, completion and cleanup remain authoritative. Operator Stop
intent is separate from physical containment. Saved legacy agents retain
their selected runtime. Prior shared title-save/deletion failures remain
disclosed.

## Model Used

OpenAI Codex, GPT-6 family, assisted with reasoning, repository tools,
code execution, browser verification, and delegated code review. The
serving model’s exact API ID and context window are not exposed. No
value is inferred from the model catalog.

## Checklist

- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used and the limits of available model
metadata
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have linked public issues or described the issue in this PR
- [x] I have not referenced internal or instance-local Paperclip issues
or links
- [x] My branch describes the change and contains no internal ticket ID
- [x] I have run the scoped local tests reported above; pending gates
are stated separately
- [x] I have added or updated tests where applicable
- [x] I have updated relevant documentation
- [x] I have considered and documented the risks
- [x] All required CI gates pass on the current head
- [x] Fresh Greptile is 5/5 with no unresolved findings on the current
head
- [x] I will address all reviewer comments before requesting merge

---------

Co-authored-by: Paperclip <noreply@paperclip.ing>
2026-11-03 09:01:39 -06:00

589 lines
40 KiB
TypeScript

import { claimNativeSessionResumptions } from "../native-runtime/native-finalization-reconciler.js";
import type { Db } from "@paperclipai/db";
import { authUsers, agents, agentWakeupRequests, caseHandoffs, cases, chatCompletionDeliveries, companies, createDb, EMBEDDED_POSTGRES_TEST_TIMEOUT_MS, heartbeatRuns, issueComments, issueThreadInteractions, issues, nativeRunFinalizations } from "@paperclipai/db";
import { eq, inArray, sql } from "drizzle-orm";
import { afterAll, beforeAll, beforeEach, describe, expect, it, vi } from "vitest";
import { resolveExternalTestDatabaseUrl } from "../../__tests__/helpers/external-test-database.js";
import { getEmbeddedPostgresTestSupport, startEmbeddedPostgresTestDatabase } from "../../__tests__/helpers/embedded-postgres.js";
import { createPostgresWakeQueueAdapter } from "../../modules/wake-queue/adapters/postgres.js";
import { createReleaseIssueExecution } from "../../modules/wake-queue/application/use-cases.js";
import { createHeartbeatQueue, type HeartbeatQueueDependencies } from "./queue.js";
import { settleConversationTurn } from "../agent-conversations.js";
import { chatCompletionDeliveryService, recordChatCompletion, recordChatHandoff } from "../chat-completion-delivery.js";
import { instanceSettingsService } from "../instance-settings.js";
import { createHeartbeatRunControl } from "./run-control.js";
import { callbacks as runControlCallbacks } from "./__tests__/run-control-fixtures.js";
import { createHeartbeatQueueTestDependencies as callbacks } from "../../__tests__/helpers/heartbeat-queue.js";
function guardedDatabase() {
const access = vi.fn(() => { throw new Error("Unexpected queue database access"); });
return { db: new Proxy({}, { get: access }) as Db, access };
}
describe("heartbeat queue module boundary", () => {
it("constructs without querying, dispatching, or registering background work", () => {
const database = guardedDatabase();
const deps = callbacks(database.db);
createHeartbeatQueue(database.db, deps);
expect(database.access).not.toHaveBeenCalled();
for (const callback of Object.values(deps)) if (vi.isMockFunction(callback)) expect(callback).not.toHaveBeenCalled();
expect(deps.activeRunExecutionPromises.size).toBe(0);
expect(deps.activeWakeupPromises.size).toBe(0);
});
it.each(["startNextQueuedRunForAgent", "resumeQueuedRuns"] as const)("honors suppression before %s reads or resumes work", async (operation) => {
const database = guardedDatabase();
const deps = callbacks(database.db);
deps.getSchedulingSuppression.mockResolvedValue({ suppressed: true, reason: "task_drain" });
const queue = createHeartbeatQueue(database.db, deps);
await queue[operation]("agent");
expect(database.access).not.toHaveBeenCalled();
expect(deps.getAgent).not.toHaveBeenCalled();
expect(deps.resumeExecutionWaitComments).not.toHaveBeenCalled();
expect(deps.executeRun).not.toHaveBeenCalled();
});
it("tracks pending wakes across instances and clears each rejected wake", async () => {
const database = guardedDatabase();
const deps = callbacks(database.db);
let rejectFirst!: (reason: Error) => void;
let rejectSecond!: (reason: Error) => void;
deps.getAgent.mockImplementationOnce(() => new Promise((_resolve, reject) => { rejectFirst = reject; }));
deps.getAgent.mockImplementationOnce(() => new Promise((_resolve, reject) => { rejectSecond = reject; }));
const first = createHeartbeatQueue(database.db, deps).trackWakeup("first");
const second = createHeartbeatQueue(database.db, deps).trackWakeup("second");
const firstCheck = expect(first).rejects.toThrow("first lookup failed");
const secondCheck = expect(second).rejects.toThrow("second lookup failed");
expect(deps.activeWakeupPromises).toEqual(new Set([first, second]));
rejectFirst(new Error("first lookup failed"));
await firstCheck;
await Promise.resolve();
expect(deps.activeWakeupPromises).toEqual(new Set([second]));
rejectSecond(new Error("second lookup failed"));
await secondCheck;
await Promise.resolve();
expect(deps.activeWakeupPromises.size).toBe(0);
expect(deps.activeRunExecutionPromises.size).toBe(0);
expect(database.access).not.toHaveBeenCalled();
});
it("uses the same policy limits for dispatch and execution", () => {
const database = guardedDatabase();
const queue = createHeartbeatQueue(database.db, callbacks(database.db));
const agent = { adapterType: "paperclip_runner", adapterConfig: { provider: "openai_dot" }, runtimeConfig: { heartbeat: { maxConcurrentRuns: 20, maxDailyRuns: 3.8 } } } as unknown as typeof agents.$inferSelect;
expect(queue.parseHeartbeatPolicy(agent)).toMatchObject({ maxConcurrentRuns: 1, maxDailyRuns: 3 });
expect(queue.parseHeartbeatPolicy({ ...agent, adapterConfig: { provider: "codex" } })).toMatchObject({ maxConcurrentRuns: 20, maxDailyRuns: 3 });
expect(database.access).not.toHaveBeenCalled();
});
});
const external = resolveExternalTestDatabaseUrl("PAPERCLIP_QUEUE_TEST_DATABASE_URL");
const support = external ? { supported: true, reason: undefined } : await getEmbeddedPostgresTestSupport();
if (!support.supported) console.warn(`Skipping queue module database tests: ${support.reason}`);
describe.skipIf(!support.supported)("heartbeat queue database wiring", () => {
let database: Awaited<ReturnType<typeof startEmbeddedPostgresTestDatabase>> | undefined;
let db: Db;
beforeAll(async () => {
if (!external) database = await startEmbeddedPostgresTestDatabase("paperclip-heartbeat-queue-module");
db = createDb(external ?? database!.connectionString);
}, EMBEDDED_POSTGRES_TEST_TIMEOUT_MS);
afterAll(async () => {
await db?.$client.end();
await database?.cleanup();
});
beforeEach(async () => {
await db.execute(sql`truncate table companies restart identity cascade`);
});
it.each([
{ provider: "codex" }, { provider: "acpx", acpxAgent: "antigravity" }, { provider: "muse" },
])("rejects direct taskless timer wakes for $provider before a run or request can be persisted", async adapterConfig => {
const [company] = await db.insert(companies).values({ name: "Timer scope", issuePrefix: "TS" }).returning();
const [agent] = await db.insert(agents).values({ companyId: company.id, name: "Ready for a task", status: "idle",
adapterType: "paperclip_runner", adapterConfig, runtimeConfig: { heartbeat: { enabled: true, intervalSec: 60 } } }).returning();
const deps = callbacks(db);
deps.getAgent.mockResolvedValue(agent);
const queue = createHeartbeatQueue(db, deps);
expect(await queue.enqueueWakeup(agent.id, { source: "timer", reason: "heartbeat_timer", requestedByActorType: "system" })).toBeNull();
expect(await db.select().from(heartbeatRuns)).toHaveLength(0);
expect(await db.select().from(agentWakeupRequests)).toHaveLength(0);
expect((await db.select().from(agents).where(eq(agents.id, agent.id)))[0])
.toMatchObject({ status: "idle", lastHeartbeatAt: null });
expect(deps.executeRun).not.toHaveBeenCalled();
expect(deps.budgets.getInvocationBlock).not.toHaveBeenCalled();
});
it("does not admit recovery after the conversation settles between precheck and admission", async () => {
const [user] = await db.insert(authUsers).values({
id: "queue-conversation-race-user", name: "Conversation owner", email: "queue-race@paperclip.test",
createdAt: new Date(), updatedAt: new Date(),
}).onConflictDoUpdate({ target: authUsers.id, set: { name: "Conversation owner" } }).returning();
const [company] = await db.insert(companies).values({ name: "Conversation race", issuePrefix: "CR" }).returning();
const [agent] = await db.insert(agents).values({ companyId: company.id, name: "Chat agent",
status: "idle", adapterType: "process", adapterConfig: {} }).returning();
const [issue] = await db.insert(issues).values({ companyId: company.id, title: "Answered conversation",
assigneeAgentId: agent.id, responsibleUserId: user.id, conversationAgentId: agent.id,
conversationUserId: user.id, conversationState: "active", status: "in_progress" }).returning();
const [completed] = await db.insert(heartbeatRuns).values({ companyId: company.id, agentId: agent.id,
status: "succeeded", finishedAt: new Date(), invocationSource: "on_demand", triggerDetail: "manual",
contextSnapshot: { issueId: issue.id, conversationSessionGeneration: issue.conversationSessionGeneration },
}).returning();
await db.insert(issueComments).values({ companyId: company.id, issueId: issue.id, authorAgentId: agent.id,
createdByRunId: completed.id, origin: "comment", body: "The requested answer." });
const deps = callbacks(db);
deps.getAgent.mockResolvedValue(agent);
deps.instanceSettings.getExperimental.mockResolvedValue({ enableAgentChat: true } as Awaited<ReturnType<typeof deps.instanceSettings.getExperimental>>);
// Admit through the normal transaction, but hold dispatch so no adapter runs.
deps.getSchedulingSuppression.mockResolvedValue({ suppressed: true, reason: "task_drain" });
deps.getSchedulingSuppression.mockResolvedValueOnce({ suppressed: false, reason: null });
const getIssueExecutionContext = deps.getIssueExecutionContext;
const outsideStates: Array<{ status: string; conversationState: string | null }> = [];
deps.getIssueExecutionContext = async (...args) => {
const current = await getIssueExecutionContext(...args);
if (current) outsideStates.push({ status: current.status, conversationState: current.conversationState });
return current;
};
let reached!: () => void;
let release!: () => void;
const atBarrier = new Promise<void>(resolve => { reached = resolve; });
const admissionMayContinue = new Promise<void>(resolve => { release = resolve; });
deps.budgets.getInvocationBlock.mockImplementationOnce(async () => {
reached();
await admissionMayContinue;
return null;
});
const pending = createHeartbeatQueue(db, deps).enqueueWakeup(agent.id, {
source: "automation", triggerDetail: "system", reason: "issue.continuation_recovery",
requestedByActorType: "system", payload: { issueId: issue.id },
contextSnapshot: { issueId: issue.id, retryOfRunId: completed.id, source: "issue.continuation_recovery" },
});
try {
await atBarrier;
expect(outsideStates).toEqual([{ status: "in_progress", conversationState: "active" }]);
expect(await settleConversationTurn(db, completed)).toBe(true);
expect((await db.select().from(issues).where(eq(issues.id, issue.id)))[0])
.toMatchObject({ status: "in_review", conversationState: "waiting" });
} finally { release(); }
const admitted = await pending;
expect(deps.executeRun).not.toHaveBeenCalled();
expect(admitted).toBeNull();
expect((await db.select().from(heartbeatRuns).where(eq(heartbeatRuns.agentId, agent.id))).map(run => run.id))
.toEqual([completed.id]);
expect(await db.select().from(agentWakeupRequests).where(eq(agentWakeupRequests.agentId, agent.id)))
.toHaveLength(0);
});
async function waitingConversationFixture() {
const [user] = await db.insert(authUsers).values({ id: "queue-conversation-owner", name: "Owner",
email: "queue-owner@paperclip.test", createdAt: new Date(), updatedAt: new Date() })
.onConflictDoUpdate({ target: authUsers.id, set: { name: "Owner" } }).returning();
const [company] = await db.insert(companies).values({ name: "Waiting conversation", issuePrefix: "WC" }).returning();
const [agent] = await db.insert(agents).values({ companyId: company.id, name: "Chat agent", status: "idle", adapterType: "process" }).returning();
const [issue] = await db.insert(issues).values({ companyId: company.id, title: "Waiting for input",
assigneeAgentId: agent.id, responsibleUserId: user.id, conversationAgentId: agent.id,
conversationUserId: user.id, conversationState: "waiting", status: "in_review" }).returning();
const deps = callbacks(db);
deps.getAgent.mockResolvedValue(agent);
deps.instanceSettings.getExperimental.mockResolvedValue({ enableAgentChat: true } as Awaited<ReturnType<typeof deps.instanceSettings.getExperimental>>);
deps.getSchedulingSuppression.mockResolvedValue({ suppressed: true, reason: "task_drain" });
deps.getSchedulingSuppression.mockResolvedValueOnce({ suppressed: false, reason: null });
await instanceSettingsService(db).updateExperimental({ enableAgentChat: true });
return { user, company, agent, issue, deps, queue: createHeartbeatQueue(db, deps) };
}
it.each(["comment", "interaction", "ordinary task"] as const)("preserves admission for a current %s", async cause => {
const f = await waitingConversationFixture();
const context: Record<string, unknown> = { issueId: f.issue.id, source: "issue.continuation_recovery" };
const payload: Record<string, unknown> = { issueId: f.issue.id };
if (cause === "comment") {
const [comment] = await db.insert(issueComments).values({ companyId: f.company.id, issueId: f.issue.id,
authorUserId: f.user.id, body: "A new question" }).returning();
context.wakeCommentId = comment.id;
payload.commentId = comment.id;
} else if (cause === "interaction") {
const [interaction] = await db.insert(issueThreadInteractions).values({ companyId: f.company.id,
issueId: f.issue.id, createdByAgentId: f.agent.id, kind: "ask_user_questions", status: "answered",
resolvedByUserId: f.user.id, resolvedAt: new Date(),
payload: { version: 1, questions: [{ id: "q", prompt: "Continue?", selectionMode: "single",
options: [{ id: "a", label: "Continue" }] }] },
result: { version: 1, answers: [{ questionId: "q", optionIds: ["a"] }] } }).returning();
Object.assign(context, { source: "issue.interaction.respond", interactionId: interaction.id,
interactionKind: interaction.kind, interactionStatus: interaction.status });
Object.assign(payload, { mutation: "interaction", interactionId: interaction.id });
} else {
await db.update(issues).set({ conversationAgentId: null, conversationUserId: null,
conversationState: null, status: "in_progress" }).where(eq(issues.id, f.issue.id));
}
const admitted = await f.queue.enqueueWakeup(f.agent.id, { source: "automation", triggerDetail: "system",
reason: cause === "ordinary task" ? "issue.continuation_recovery" : "issue_commented",
requestedByActorType: "system", payload, contextSnapshot: context });
expect(admitted).toMatchObject({ agentId: f.agent.id, status: "queued" });
expect(f.deps.executeRun).not.toHaveBeenCalled();
expect(await db.select().from(agentWakeupRequests).where(eq(agentWakeupRequests.agentId, f.agent.id)))
.toHaveLength(1);
});
it("delivers a real task completion to its waiting conversation", async () => {
const f = await waitingConversationFixture();
const [source] = await db.insert(heartbeatRuns).values({ companyId: f.company.id, agentId: f.agent.id,
status: "succeeded", finishedAt: new Date(), contextSnapshot: { issueId: f.issue.id, conversationSessionGeneration: 0 } }).returning();
const [task] = await db.insert(issues).values({ companyId: f.company.id, title: "Handed-off task",
status: "in_progress", createdByAgentId: f.agent.id }).returning();
await recordChatHandoff(db, task, source.id);
const recorded = await db.transaction(async tx => {
const [done] = await tx.update(issues).set({ status: "done" }).where(eq(issues.id, task.id)).returning();
return recordChatCompletion(tx, task, done);
});
expect(recorded).toBe(true);
await chatCompletionDeliveryService(db, { wakeup: (agentId, options) => f.queue.enqueueWakeup(agentId, options) }).sweepPending();
const [delivery] = await db.select().from(chatCompletionDeliveries).where(eq(chatCompletionDeliveries.companyId, f.company.id));
expect(delivery).toMatchObject({ status: "queued" });
const [run] = await db.select().from(heartbeatRuns).where(eq(heartbeatRuns.id, delivery.targetRunId!));
expect(run).toMatchObject({ agentId: f.agent.id, status: "queued" });
expect(run.contextSnapshot).toMatchObject({ issueId: f.issue.id, wakeReason: "chat_task_completed" });
expect(f.deps.executeRun).not.toHaveBeenCalled();
});
async function fixture(status = "running") {
const [company] = await db.insert(companies).values({ name: "Queue module", issuePrefix: "QM" }).returning();
const [agent] = await db.insert(agents).values({ companyId: company.id, name: "Queue agent", role: "engineer", status: "idle", adapterType: "process", adapterConfig: {} }).returning();
const [issue] = await db.insert(issues).values({ companyId: company.id, title: "Queue task", status: "todo", assigneeAgentId: agent.id }).returning();
const [wake] = await db.insert(agentWakeupRequests).values({ companyId: company.id, agentId: agent.id, source: "assignment", status: "claimed", claimedAt: new Date() }).returning();
const [run] = await db.insert(heartbeatRuns).values({ companyId: company.id, agentId: agent.id, status, invocationSource: "assignment", wakeupRequestId: wake.id, startedAt: status === "running" ? new Date() : null, contextSnapshot: { issueId: issue.id } }).returning();
await db.update(issues).set({ executionRunId: run.id, executionAgentNameKey: agent.name, executionLockedAt: new Date() }).where(eq(issues.id, issue.id));
return { company, agent, issue, wake, run };
}
async function strandedMention(runtimeMode = "legacy") {
const seeded = await fixture("succeeded");
await db.update(heartbeatRuns).set({ runtimeMode, nativeIssueId: runtimeMode === "native" ? seeded.issue.id : null, finishedAt: new Date() }).where(eq(heartbeatRuns.id, seeded.run.id));
if (runtimeMode === "native") await db.insert(nativeRunFinalizations).values({ runId: seeded.run.id, companyId: seeded.company.id, issueId: seeded.issue.id, phase: "committed" });
await db.update(issues).set({ executionRunId: null, checkoutRunId: null }).where(eq(issues.id, seeded.issue.id));
const [recipient] = await db.insert(agents).values({ companyId: seeded.company.id, name: "Mention recipient", role: "engineer", status: "idle", adapterType: "process", adapterConfig: {} }).returning();
const [comment] = await db.insert(issueComments).values({ companyId: seeded.company.id, issueId: seeded.issue.id, body: "Saved feedback", authorAgentId: seeded.agent.id }).returning();
const [pending] = await db.insert(agentWakeupRequests).values({ companyId: seeded.company.id, agentId: recipient.id,
source: "automation", reason: "issue_execution_deferred", status: "deferred_issue_execution",
payload: { issueId: seeded.issue.id, _paperclipWakeContext: { issueId: seeded.issue.id, wakeReason: "issue_comment_mentioned", wakeCommentIds: [comment.id] } },
}).returning();
return { ...seeded, recipient, pending };
}
async function addStrandedMessages(seeded: Awaited<ReturnType<typeof strandedMention>>) {
const messages = [seeded.pending];
for (let index = 0; index < 2; index++) {
const [comment] = await db.insert(issueComments).values({ companyId: seeded.company.id, issueId: seeded.issue.id,
body: `Additional saved feedback ${index}`, authorAgentId: seeded.agent.id }).returning();
const [wake] = await db.insert(agentWakeupRequests).values({ companyId: seeded.company.id, agentId: seeded.recipient.id,
source: "automation", reason: "issue_execution_deferred", status: "deferred_issue_execution",
payload: { issueId: seeded.issue.id, _paperclipWakeContext: { issueId: seeded.issue.id,
wakeReason: "issue_comment_mentioned", wakeCommentIds: [comment.id] } } }).returning();
messages.push(wake);
}
await db.update(agentWakeupRequests).set({ updatedAt: new Date(0) }).where(inArray(agentWakeupRequests.id, messages.map(row => row.id)));
return messages;
}
async function expectSavedMessages(messages: Awaited<ReturnType<typeof addStrandedMessages>>) {
const rows = await db.select().from(agentWakeupRequests).where(inArray(agentWakeupRequests.id, messages.map(row => row.id)));
expect(rows).toHaveLength(messages.length);
for (const before of messages) {
const row = rows.find(row => row.id === before.id)!;
expect(row).toMatchObject({ payload: before.payload, status: "deferred_issue_execution", runId: null,
finishedAt: null, error: null });
expect(row.updatedAt.getTime()).toBeGreaterThan(0);
}
}
it.each(["legacy", "native"])("keeps saved %s messages while avoiding repeated native source release", async runtimeMode => {
const seeded = await strandedMention(runtimeMode);
const messages = await addStrandedMessages(seeded);
const deps = callbacks(db), queue = createHeartbeatQueue(db, deps);
await queue.resumeQueuedRuns();
expect(deps.releaseIssueExecutionAndPromote).toHaveBeenCalledTimes(runtimeMode === "native" ? 1 : 3);
await expectSavedMessages(messages);
await queue.resumeQueuedRuns();
expect(deps.releaseIssueExecutionAndPromote).toHaveBeenCalledTimes(runtimeMode === "native" ? 2 : 6);
expect(deps.executeRun).not.toHaveBeenCalled();
});
it("retries a rejected native release on the next saved message", async () => {
const seeded = await strandedMention("native"), messages = await addStrandedMessages(seeded);
const deps = callbacks(db);
deps.releaseIssueExecutionAndPromote.mockRejectedValueOnce(new Error("Fixture release remains pending"));
await createHeartbeatQueue(db, deps).resumeQueuedRuns();
expect(deps.releaseIssueExecutionAndPromote).toHaveBeenCalledTimes(2);
await expectSavedMessages(messages);
expect(deps.executeRun).not.toHaveBeenCalled();
});
it("joins the native release before it suppresses another message's duplicate check", async () => {
const seeded = await strandedMention("native"), messages = await addStrandedMessages(seeded);
const deps = callbacks(db);
let finish!: () => void;
deps.releaseIssueExecutionAndPromote.mockImplementationOnce(() => new Promise<void>(resolve => { finish = resolve; }));
let settled = false;
const pending = createHeartbeatQueue(db, deps).resumeQueuedRuns().then(() => { settled = true; });
try {
await vi.waitFor(() => expect(deps.releaseIssueExecutionAndPromote).toHaveBeenCalledTimes(1));
expect(settled).toBe(false);
} finally { finish?.(); }
await pending;
expect(deps.releaseIssueExecutionAndPromote).toHaveBeenCalledTimes(1);
await expectSavedMessages(messages);
});
it("rechecks a different native source that completes during the pass", async () => {
const seeded = await strandedMention("native"), messages = await addStrandedMessages(seeded);
await db.update(heartbeatRuns).set({ createdAt: new Date(0) }).where(eq(heartbeatRuns.id, seeded.run.id));
const deps = callbacks(db);
let nextId: string | undefined;
deps.releaseIssueExecutionAndPromote.mockImplementationOnce(async () => {
const [next] = await db.insert(heartbeatRuns).values({ companyId: seeded.company.id, agentId: seeded.agent.id,
runtimeMode: "native", nativeIssueId: seeded.issue.id, status: "succeeded", finishedAt: new Date(),
contextSnapshot: { issueId: seeded.issue.id } }).returning();
nextId = next.id;
});
await createHeartbeatQueue(db, deps).resumeQueuedRuns();
expect(deps.releaseIssueExecutionAndPromote.mock.calls.map(([run]) => run.id)).toEqual([seeded.run.id, nextId]);
await expectSavedMessages(messages);
});
it("rechecks changed native evidence on the same source run", async () => {
const seeded = await strandedMention("native"), messages = await addStrandedMessages(seeded);
const deps = callbacks(db);
deps.releaseIssueExecutionAndPromote.mockImplementationOnce(async () => {
await db.update(heartbeatRuns).set({ resultJson: { observedReceipt: "new evidence" } }).where(eq(heartbeatRuns.id, seeded.run.id));
});
await createHeartbeatQueue(db, deps).resumeQueuedRuns();
expect(deps.releaseIssueExecutionAndPromote).toHaveBeenCalledTimes(2);
expect(deps.releaseIssueExecutionAndPromote.mock.calls.map(([run]) => run.id)).toEqual([seeded.run.id, seeded.run.id]);
await expectSavedMessages(messages);
});
it("checks each unrelated native task in the same pass", async () => {
const seeded = await strandedMention("native"), firstMessages = await addStrandedMessages(seeded);
const [issue] = await db.insert(issues).values({ companyId: seeded.company.id, title: "Another saved task", status: "todo", assigneeAgentId: seeded.agent.id }).returning();
const [run] = await db.insert(heartbeatRuns).values({ companyId: seeded.company.id, agentId: seeded.agent.id,
runtimeMode: "native", nativeIssueId: issue.id, status: "succeeded", finishedAt: new Date(), contextSnapshot: { issueId: issue.id } }).returning();
const [comment] = await db.insert(issueComments).values({ companyId: seeded.company.id, issueId: issue.id,
body: "Other task feedback", authorAgentId: seeded.agent.id }).returning();
const [pending] = await db.insert(agentWakeupRequests).values({ companyId: seeded.company.id, agentId: seeded.recipient.id,
source: "automation", reason: "issue_execution_deferred", status: "deferred_issue_execution",
payload: { issueId: issue.id, _paperclipWakeContext: { issueId: issue.id, wakeReason: "issue_comment_mentioned", wakeCommentIds: [comment.id] } } }).returning();
const secondMessages = await addStrandedMessages({ ...seeded, issue, run, pending });
const deps = callbacks(db);
await createHeartbeatQueue(db, deps).resumeQueuedRuns();
expect(deps.releaseIssueExecutionAndPromote).toHaveBeenCalledTimes(2);
expect(new Set(deps.releaseIssueExecutionAndPromote.mock.calls.map(([row]) => row.id))).toEqual(new Set([seeded.run.id, run.id]));
await expectSavedMessages([...firstMessages, ...secondMessages]);
});
it("rechecks a handoff's assignment version under the issue claim lock", async () => {
const { company, agent, issue, wake, run } = await fixture("queued");
const [record] = await db.insert(cases).values({ companyId: company.id, caseNumber: 1, identifier: "QM-CASE-1",
caseType: "work", title: issue.title, workIssueId: issue.id, handoffVersion: 1 }).returning();
await db.insert(caseHandoffs).values({ companyId: company.id, caseId: record.id, issueId: issue.id,
version: 1, toAgentId: agent.id, phase: "implementation", instructions: "Build the change", actorType: "system",
actorId: "test", idempotencyKey: "queue-version-race", inputFingerprint: "fixture", wakeupRequestId: wake.id,
expectedIssueStatusVersion: issue.statusVersion, appliedIssueStatusVersion: issue.statusVersion, state: "delivered" });
const deps = callbacks(db);
deps.getAgent.mockResolvedValue(agent);
deps.instanceSettings.getExperimental.mockResolvedValue({ enableCases: true } as Awaited<ReturnType<HeartbeatQueueDependencies["instanceSettings"]["getExperimental"]>>);
// Simulate A → B → A after the optimistic receipt read. The agent ID is
// unchanged, but the original assignment no longer authorizes a new turn.
deps.budgets.getInvocationBlock.mockImplementationOnce(async () => {
await db.update(issues).set({ statusVersion: issue.statusVersion + 2 }).where(eq(issues.id, issue.id));
return null;
});
expect(await createHeartbeatQueue(db, deps).claimQueuedRun({ ...run, responsibleUserId: "queue-test-user" })).toBeNull();
expect((await db.select().from(heartbeatRuns).where(eq(heartbeatRuns.id, run.id)))[0])
.toMatchObject({ status: "queued", startedAt: null });
expect(deps.executeRun).not.toHaveBeenCalled();
});
it.each(["legacy", "native"])("periodic recovery preserves a paused recipient through the real %s release adapter", async runtimeMode => {
const seeded = await strandedMention(runtimeMode);
const { company, issue, recipient } = seeded;
const messages = await addStrandedMessages(seeded);
await db.update(agents).set({ status: "paused" }).where(eq(agents.id, recipient.id));
await db.update(issues).set({ status: "blocked" }).where(eq(issues.id, issue.id));
const deps = callbacks(db);
const release = createReleaseIssueExecution({ issueLock: createPostgresWakeQueueAdapter(db, {
resolveResponsibleUserId: async () => "responsible-user",
getRoutineEnv: async () => ({ routineId: null, env: null, responsibleUserId: null }),
resolveSessionBeforeForWakeup: async () => null,
}), recovery: { escalateStrandedAssignedIssue: async () => undefined, escalateStrandedRecoveryIssueInPlace: async () => undefined } });
const controlDeps = runControlCallbacks(db);
vi.mocked(controlDeps.wakeQueue.releaseIssueExecution).mockImplementation(release);
const control = createHeartbeatRunControl(db, controlDeps);
deps.releaseIssueExecutionAndPromote.mockImplementation(control.releaseIssueExecutionAndPromote);
for (let pass = 0; pass < 2; pass++) {
await createHeartbeatQueue(db, deps).resumeQueuedRuns();
await expectSavedMessages(messages);
expect(await db.select().from(heartbeatRuns).where(eq(heartbeatRuns.companyId, company.id))).toHaveLength(1);
}
expect(deps.releaseIssueExecutionAndPromote).toHaveBeenCalledTimes(runtimeMode === "native" ? 2 : 6);
expect(deps.executeRun).not.toHaveBeenCalled();
});
it.each(["legacy", "native"])("resumes an accepted non-assignee queue through its %s owner", async runtimeMode => {
const { run, pending } = await strandedMention(runtimeMode);
await db.update(heartbeatRuns).set({ contextSnapshot: { ...run.contextSnapshot, prompt: "large context".repeat(20_000) },
runnerProfileJson: { instructions: "large profile".repeat(20_000) }, stdoutExcerpt: "large output".repeat(20_000),
}).where(eq(heartbeatRuns.id, run.id));
const deps = callbacks(db);
await createHeartbeatQueue(db, deps).resumeQueuedRuns();
expect(deps.releaseIssueExecutionAndPromote).toHaveBeenCalledExactlyOnceWith(
expect.objectContaining({ id: run.id }), { suppressImmediateRecovery: true, resumeStrandedQueue: true },
);
expect((await db.select().from(agentWakeupRequests).where(eq(agentWakeupRequests.id, pending.id)))[0]).toMatchObject({
status: "deferred_issue_execution", runId: null, payload: pending.payload,
});
expect(deps.executeRun).not.toHaveBeenCalled();
const selected = deps.releaseIssueExecutionAndPromote.mock.calls[0][0];
expect(Buffer.byteLength(JSON.stringify(selected))).toBeLessThan(512);
expect(selected).not.toHaveProperty("contextSnapshot");
expect(selected).not.toHaveProperty("runnerProfileJson");
expect(selected).not.toHaveProperty("stdoutExcerpt");
});
it("preserves a former owner's ordinary messages for the assignment handoff", async () => {
const { pending } = await strandedMention();
const payload = { ...pending.payload, _paperclipWakeContext: {
...(pending.payload!._paperclipWakeContext as Record<string, unknown>), wakeReason: "issue_commented",
} };
await db.update(agentWakeupRequests).set({ payload }).where(eq(agentWakeupRequests.id, pending.id));
const deps = callbacks(db);
await createHeartbeatQueue(db, deps).resumeQueuedRuns();
expect(deps.releaseIssueExecutionAndPromote).not.toHaveBeenCalled();
expect((await db.select().from(agentWakeupRequests).where(eq(agentWakeupRequests.id, pending.id)))[0])
.toMatchObject({ status: "deferred_issue_execution", runId: null, payload });
expect(deps.executeRun).not.toHaveBeenCalled();
});
it.each(["queued", "running", "scheduled_retry"])("does not replay a finishing owner while another agent has a %s turn", async status => {
const { company, issue, recipient } = await strandedMention();
await db.insert(heartbeatRuns).values({ companyId: company.id, agentId: recipient.id, status, invocationSource: "assignment", contextSnapshot: { issueId: issue.id } });
const deps = callbacks(db);
deps.getAgent.mockResolvedValue({ ...recipient, status: "paused" });
await createHeartbeatQueue(db, deps).resumeQueuedRuns();
expect(deps.releaseIssueExecutionAndPromote).not.toHaveBeenCalled();
expect(deps.executeRun).not.toHaveBeenCalled();
});
it.each(["claimed", "cancelled"] as const)("releases a running claim while preserving a %s wake", async (wakeStatus) => {
const { run, issue, wake } = await fixture();
await db.update(agentWakeupRequests).set({ status: wakeStatus }).where(eq(agentWakeupRequests.id, wake.id));
await createHeartbeatQueue(db, callbacks(db)).releaseRunClaimedJustBeforeSuppression(run.id);
const [released] = await db.select().from(heartbeatRuns).where(eq(heartbeatRuns.id, run.id));
const [releasedIssue] = await db.select().from(issues).where(eq(issues.id, issue.id));
const [releasedWake] = await db.select().from(agentWakeupRequests).where(eq(agentWakeupRequests.id, wake.id));
expect(released).toMatchObject({ status: "queued", startedAt: null, responsibleUserId: null });
expect(releasedIssue).toMatchObject({ executionRunId: null, executionAgentNameKey: null, executionLockedAt: null });
expect(releasedWake.status).toBe(wakeStatus === "cancelled" ? "cancelled" : "queued");
if (wakeStatus === "claimed") expect(releasedWake.claimedAt).toBeNull();
});
it("preserves a newer execution owner when releasing an old claim", async () => {
const { run, issue, agent, company } = await fixture();
const [newOwner] = await db.insert(heartbeatRuns).values({ companyId: company.id, agentId: agent.id, status: "running", invocationSource: "assignment" }).returning();
await db.update(issues).set({ executionRunId: newOwner.id }).where(eq(issues.id, issue.id));
await createHeartbeatQueue(db, callbacks(db)).releaseRunClaimedJustBeforeSuppression(run.id);
const [current] = await db.select().from(issues).where(eq(issues.id, issue.id));
expect(current.executionRunId).toBe(newOwner.id);
});
it.each(["running", "failed"])("resumes the %s native owner before promoting its queued waiter", async status => {
const { run, issue, agent, company } = await fixture(status);
await db.update(heartbeatRuns).set({ runtimeMode: "native", nativeIssueId: issue.id, runnerProfileJson: {
nativeExecutionInput: { binding: { runId: run.id } },
sessionCheckpoint: { providerSessionId: "original-session" },
} }).where(eq(heartbeatRuns.id, run.id));
await db.insert(nativeRunFinalizations).values({ runId: run.id, companyId: company.id,
issueId: issue.id, phase: "retryable_failure", attempt: 1 });
const [waiter] = await db.insert(heartbeatRuns).values({ companyId: company.id, agentId: agent.id,
status: "queued", invocationSource: "assignment", responsibleUserId: "queue-test-user", contextSnapshot: { issueId: issue.id },
}).returning();
const deps = callbacks(db);
deps.getAgent.mockResolvedValue(agent);
deps.instanceSettings.getExperimental.mockResolvedValue({ enableNativeRunner: true } as Awaited<ReturnType<HeartbeatQueueDependencies["instanceSettings"]["getExperimental"]>>);
const queue = createHeartbeatQueue(db, deps);
expect(await queue.claimQueuedRun(waiter)).toBeNull();
const claims = await claimNativeSessionResumptions({ db, runnerInstanceId: "owner", runIds: [run.id] });
expect(claims.map(claim => claim.runId)).toEqual([run.id]);
expect(await queue.claimQueuedRun(waiter)).toBeNull();
await db.update(heartbeatRuns).set({ status: "succeeded" }).where(eq(heartbeatRuns.id, run.id));
await db.update(nativeRunFinalizations).set({ phase: "committed", leaseOwner: null, leaseExpiresAt: null })
.where(eq(nativeRunFinalizations.runId, run.id));
await db.update(issues).set({ executionRunId: null }).where(eq(issues.id, issue.id));
expect(await queue.claimQueuedRun(waiter)).toMatchObject({ id: waiter.id, status: "running" });
expect((await db.select().from(issues).where(eq(issues.id, issue.id)))[0].executionRunId).toBe(waiter.id);
});
it.each(["direct", "batch"] as const)("keeps fresh work queued across a held external connection on the %s path", async path => {
const { run, agent, issue, wake } = await fixture("queued");
await db.update(heartbeatRuns).set({ responsibleUserId: "queue-test-user" }).where(eq(heartbeatRuns.id, run.id));
const deps = { ...callbacks(db), canStartNewExternalBotWork: vi.fn(async () => false) };
deps.getAgent.mockResolvedValue({ ...agent, adapterType: "paperclip_runner", adapterConfig: { provider: "grok_bot" } });
const queue = createHeartbeatQueue(db, deps);
for (let attempt = 0; attempt < 3; attempt++) {
if (path === "direct") expect(await queue.claimQueuedRun({ ...run, responsibleUserId: "queue-test-user" })).toBeNull();
else expect(await queue.startNextQueuedRunForAgent(agent.id)).toEqual([]);
}
expect((await db.select().from(heartbeatRuns).where(eq(heartbeatRuns.id, run.id)))[0]).toMatchObject({ status: "queued", startedAt: null });
expect((await db.select().from(agentWakeupRequests).where(eq(agentWakeupRequests.id, wake.id)))[0].status).toBe(wake.status);
expect((await db.select().from(issues).where(eq(issues.id, issue.id)))[0].status).toBe("todo");
expect(deps.executeRun).not.toHaveBeenCalled();
expect(deps.setRunStatus).not.toHaveBeenCalled();
});
it("admits the same queued input after the external connection is repaired", async () => {
const { run, agent } = await fixture("queued");
await db.update(heartbeatRuns).set({ responsibleUserId: "queue-test-user" }).where(eq(heartbeatRuns.id, run.id));
const deps = { ...callbacks(db), canStartNewExternalBotWork: vi.fn(async () => false) };
deps.getAgent.mockResolvedValue(agent);
const queue = createHeartbeatQueue(db, deps);
expect(await queue.startNextQueuedRunForAgent(agent.id)).toEqual([]);
deps.canStartNewExternalBotWork.mockResolvedValue(true);
expect(await queue.startNextQueuedRunForAgent(agent.id)).toEqual([expect.objectContaining({ id: run.id, status: "running" })]);
expect(deps.executeRun).toHaveBeenCalledExactlyOnceWith(run.id);
});
it("rechecks the connection hold when Stop races the queue scan", async () => {
const { run, agent } = await fixture("queued");
const deps = { ...callbacks(db), canStartNewExternalBotWork: vi.fn(async () => false) };
deps.getAgent.mockResolvedValue(agent);
deps.canStartNewExternalBotWork.mockResolvedValueOnce(true);
expect(await createHeartbeatQueue(db, deps).startNextQueuedRunForAgent(agent.id)).toEqual([]);
expect((await db.select().from(heartbeatRuns).where(eq(heartbeatRuns.id, run.id)))[0]).toMatchObject({ status: "queued", startedAt: null });
expect(deps.canStartNewExternalBotWork).toHaveBeenCalledTimes(2);
expect(deps.executeRun).not.toHaveBeenCalled();
});
it("keeps admitted resumptions outside the fresh external connection gate", async () => {
const { run, agent } = await fixture("queued");
await db.update(heartbeatRuns).set({ startedAt: new Date(), responsibleUserId: "queue-test-user" }).where(eq(heartbeatRuns.id, run.id));
const [admitted] = await db.select().from(heartbeatRuns).where(eq(heartbeatRuns.id, run.id));
const deps = { ...callbacks(db), canStartNewExternalBotWork: vi.fn(async () => false) };
deps.getAgent.mockResolvedValue(agent);
expect(await createHeartbeatQueue(db, deps).claimQueuedRun(admitted)).toMatchObject({ id: run.id, status: "running" });
expect(deps.canStartNewExternalBotWork).not.toHaveBeenCalled();
});
it("cancels at the daily run cap before claiming or dispatching", async () => {
const { run, agent } = await fixture("queued");
const cappedAgent = { ...agent, runtimeConfig: { heartbeat: { maxDailyRuns: 0 } } };
const deps = callbacks(db);
deps.getAgent.mockResolvedValue(cappedAgent);
expect(await createHeartbeatQueue(db, deps).claimQueuedRun(run)).toBeNull();
const [cancelled] = await db.select().from(heartbeatRuns).where(eq(heartbeatRuns.id, run.id));
expect(cancelled).toMatchObject({ status: "cancelled", errorCode: "heartbeat.daily_run_limit" });
expect(deps.setWakeupStatus).toHaveBeenCalledWith(run.wakeupRequestId, "skipped", expect.any(Object));
expect(deps.releaseIssueExecutionAndPromote).toHaveBeenCalledWith(cancelled, { suppressImmediateRecovery: true });
expect(deps.treeControlSvc.getActivePauseHoldGate).not.toHaveBeenCalled();
expect(deps.executeRun).not.toHaveBeenCalled();
});
});