@oneuptime/common 14.0.9 → 14.0.10

package/Tests/Server/Utils/AI/RemediationExecutionRunnerResourceMode.test.ts

@@ -0,0 +1,1249 @@
+import RemediationExecutionRunner, {
+  ResourceBreakerState,
+  ResourceModeResolution,
+  isClusterRemediationRound,
+  isResourceRemediationRound,
+} from "../../../../Server/Utils/AI/Remediation/RemediationExecutionRunner";
+import AIInvestigationEngine, {
+  InvestigationRequest,
+} from "../../../../Server/Utils/AI/SRE/AIInvestigationEngine";
+import { ConfidenceSignal } from "../../../../Server/Utils/AI/SRE/ConfidenceSignal";
+import {
+  ObservabilityAssistantExtraTool,
+  ObservabilityAssistantResult,
+} from "../../../../Server/Utils/AI/Chat/ObservabilityAssistant";
+import { ToolCallOutcome } from "../../../../Server/Utils/AI/Toolbox/Index";
+import AIRunService from "../../../../Server/Services/AIRunService";
+import AlertFeedService from "../../../../Server/Services/AlertFeedService";
+import AutoRemediationRuleService from "../../../../Server/Services/AutoRemediationRuleService";
+import AutoRemediationSuggestionService from "../../../../Server/Services/AutoRemediationSuggestionService";
+import IncidentFeedService from "../../../../Server/Services/IncidentFeedService";
+import IncidentService from "../../../../Server/Services/IncidentService";
+import KubernetesClusterAiAccessService from "../../../../Server/Services/KubernetesClusterAiAccessService";
+import ProjectService from "../../../../Server/Services/ProjectService";
+import ResourceAiAccessService from "../../../../Server/Services/ResourceAiAccessService";
+import RunnerJobService from "../../../../Server/Services/RunnerJobService";
+import { MAX_AUTO_EXECUTIONS_PER_RULE_PER_HOUR } from "../../../../Server/Services/AutoRemediationRuleEngineService";
+import PostedRootCause from "../../../../Server/Utils/AI/SRE/PostedRootCause";
+import logger from "../../../../Server/Utils/Logger";
+import Semaphore, {
+  SemaphoreMutex,
+} from "../../../../Server/Infrastructure/Semaphore";
+import AutoRemediationSuggestion from "../../../../Models/DatabaseModels/AutoRemediationSuggestion";
+import Incident from "../../../../Models/DatabaseModels/Incident";
+import Project from "../../../../Models/DatabaseModels/Project";
+import RunnerJob from "../../../../Models/DatabaseModels/RunnerJob";
+import AutoRemediationExecutionMode from "../../../../Types/AutoRemediation/AutoRemediationExecutionMode";
+import AutoRemediationSuggestionStatus from "../../../../Types/AutoRemediation/AutoRemediationSuggestionStatus";
+import AutoRemediationSuggestionType from "../../../../Types/AutoRemediation/AutoRemediationSuggestionType";
+import AutoRemediationVerificationStatus from "../../../../Types/AutoRemediation/AutoRemediationVerificationStatus";
+import {
+  AiRemediationCommandExecutionStatus,
+  AiRemediationCommandPlan,
+  AiRemediationCommandPlanUtil,
+  AiRemediationCommandPolicyVerdict,
+  AiRemediationRollbackStatus,
+  RESOURCE_UNATTENDED_ROUND_BECOMES_PROPOSAL_SUMMARY,
+} from "../../../../Types/AutoRemediation/AiRemediationCommandPlan";
+import AiResourceType from "../../../../Types/ResourceAiAgent/AiResourceType";
+import {
+  ResourceAiAccessStatus,
+  ResourceAiAgentPosture,
+  ResourceAiRemediationMode,
+  ResourceCommandTier,
+} from "../../../../Types/ResourceAiAgent/ResourceAiAccess";
+import RunbookStepType from "../../../../Types/Runbook/RunbookStepType";
+import RunnerJobStatus from "../../../../Types/Runbook/RunnerJobStatus";
+import {
+  LIST_INFRASTRUCTURE_ACCESS_TOOL_NAME,
+  RUN_INFRASTRUCTURE_COMMAND_TOOL_NAME,
+} from "../../../../Server/Utils/AI/ResourceAccess/ResourceAccessToolNames";
+import { JSONObject } from "../../../../Types/JSON";
+import ObjectID from "../../../../Types/ObjectID";
+import PositiveNumber from "../../../../Types/PositiveNumber";
+import OneUptimeDate from "../../../../Types/Date";
+import { afterEach, beforeEach, describe, expect, it } from "@jest/globals";
+
+/*
+ * Contract under test — a RESOURCE round of RemediationExecutionRunner
+ * (remediation a resource's AI agent page asked for: a Docker or Podman
+ * host, a Docker Swarm, Proxmox, VMware or Ceph cluster, a database server,
+ * a host), built beside the cluster round and mirroring it:
+ *
+ * - the round is read off the suggestion row (a resource, no rule); its
+ *   consent is the resource's own, so the project's AI command execution
+ *   opt-in does not gate it, while Enable AI / auto-remediation still do;
+ * - the resource is re-read at the start and must be remediation-ready; a
+ *   deleted or no-longer-ready resource settles NoneApplicable with why;
+ * - the mode is resolved exactly like a cluster's: a FullAuto snapshot runs
+ *   unattended only while the resource's mode still is unattended, its
+ *   hourly breaker has headroom and no other round holds it — each
+ *   downgrade corrects the row, posts to the feed and opens the rationale;
+ * - the model gets the resource toolkit (ResourceCommand on this resource
+ *   only) plus the read-only infrastructure tool, a persona and a question
+ *   that name the resource and its programs, and a context block for it;
+ * - an unattended round that ran nothing but kept changes for a human
+ *   settles Suggested with exactly those — after re-checking the resource
+ *   live (ready, same agent, the agent's write scope) — else NoneApplicable;
+ * - earlier rounds on the same resource for the same signal are shown.
+ */
+
+const PROJECT_ID: ObjectID = new ObjectID(
+  "22222222-2222-4222-8222-222222222222",
+);
+const RUN_ID: ObjectID = new ObjectID("88888888-8888-4888-8888-888888888888");
+const SUGGESTION_ID: ObjectID = new ObjectID(
+  "77777777-7777-4777-8777-777777777777",
+);
+const OTHER_SUGGESTION_ID: ObjectID = new ObjectID(
+  "79797979-7979-4979-8979-797979797979",
+);
+const INCIDENT_ID: ObjectID = new ObjectID(
+  "cccccccc-cccc-4ccc-8ccc-cccccccccccc",
+);
+const RESOURCE_ID: ObjectID = new ObjectID(
+  "33333333-3333-4333-8333-333333333333",
+);
+const AGENT_ID: string = "44444444-4444-4444-8444-444444444444";
+const OTHER_AGENT_ID: string = "45454545-4545-4545-8545-454545454545";
+const JOB_ID: ObjectID = new ObjectID("66666666-6666-4666-8666-666666666666");
+
+function posture(
+  overrides: Partial<ResourceAiAgentPosture> = {},
+): ResourceAiAgentPosture {
+  return {
+    resourceType: AiResourceType.DockerHost,
+    resourceIdentifier: "web-1",
+    allowWrites: true,
+    writeTargets: [],
+    protectedTargets: [],
+    reachable: true,
+    ...overrides,
+  };
+}
+
+function resource(
+  overrides: Partial<ResourceAiAccessStatus> = {},
+  postureOverrides: Partial<ResourceAiAgentPosture> = {},
+): ResourceAiAccessStatus {
+  return {
+    resourceType: AiResourceType.DockerHost,
+    resourceId: RESOURCE_ID.toString(),
+    resourceName: "web-1",
+    isAiInvestigationEnabled: true,
+    aiRemediationMode: ResourceAiRemediationMode.Automatic,
+    aiCommandAllowlist: [],
+    agent: {
+      agentId: AGENT_ID,
+      connectionStatus: "connected",
+      isOnline: true,
+      posture: posture(postureOverrides),
+    },
+    gaps: [],
+    isInvestigationReady: true,
+    isRemediationReady: true,
+    ...overrides,
+  };
+}
+
+// The suggestion row as the database holds it: a resource round.
+function resourceRow(
+  overrides: Record<string, unknown> = {},
+): Record<string, unknown> {
+  return {
+    id: SUGGESTION_ID,
+    _id: SUGGESTION_ID.toString(),
+    projectId: PROJECT_ID,
+    status: AutoRemediationSuggestionStatus.Planning,
+    suggestionType: AutoRemediationSuggestionType.CommandPlan,
+    executionMode: AutoRemediationExecutionMode.FullAuto,
+    autoResolveOnRecovery: true,
+    incidentId: INCIDENT_ID,
+    resourceType: AiResourceType.DockerHost,
+    resourceId: RESOURCE_ID,
+    ruleNameSnapshot: 'AI remediation for Docker host "web-1"',
+    verificationWindowMinutes: 15,
+    createdAt: OneUptimeDate.getSomeMinutesAgo(1),
+    ...overrides,
+  };
+}
+
+function mockSuggestionHonouringSelect(
+  row: Record<string, unknown>,
+): jest.SpyInstance {
+  return jest
+    .spyOn(AutoRemediationSuggestionService, "findOneById")
+    .mockImplementation(
+      async (args: unknown): Promise<AutoRemediationSuggestion> => {
+        const select: Record<string, unknown> = (
+          args as { select: Record<string, unknown> }
+        ).select;
+        const picked: Record<string, unknown> = { id: row["id"] };
+        for (const key of Object.keys(select)) {
+          if (key in row) {
+            picked[key] = row[key];
+          }
+        }
+        return picked as unknown as AutoRemediationSuggestion;
+      },
+    );
+}
+
+function captureRequest(): { get: () => InvestigationRequest } {
+  const captured: { value: InvestigationRequest | null } = { value: null };
+
+  jest
+    .spyOn(AIInvestigationEngine, "executeRun")
+    .mockImplementation(
+      async (data: { request: InvestigationRequest }): Promise<void> => {
+        captured.value = data.request;
+      },
+    );
+
+  return {
+    get: (): InvestigationRequest => {
+      expect(captured.value).not.toBeNull();
+      return captured.value!;
+    },
+  };
+}
+
+function toolNames(request: InvestigationRequest): Array<string> {
+  return (request.extraTools || []).map(
+    (tool: ObservabilityAssistantExtraTool) => {
+      return tool.definition.name;
+    },
+  );
+}
+
+function findTool(
+  request: InvestigationRequest,
+  name: string,
+): ObservabilityAssistantExtraTool {
+  const tool: ObservabilityAssistantExtraTool | undefined = (
+    request.extraTools || []
+  ).find((candidate: ObservabilityAssistantExtraTool) => {
+    return candidate.definition.name === name;
+  });
+  expect(tool).toBeDefined();
+  return tool!;
+}
+
+function postAnalysisArgs(analysisMarkdown: string): {
+  analysisMarkdown: string;
+  confidence: ConfidenceSignal;
+  result: ObservabilityAssistantResult;
+} {
+  return {
+    analysisMarkdown,
+    confidence: {
+      confident: true,
+      source: "classification",
+    } as ConfidenceSignal,
+    result: {} as ObservabilityAssistantResult,
+  };
+}
+
+async function run(): Promise<void> {
+  await RemediationExecutionRunner.executeRemediation({
+    aiRunId: RUN_ID,
+    projectId: PROJECT_ID,
+    suggestionId: SUGGESTION_ID,
+    attemptCount: 1,
+  });
+}
+
+function resourceArgs(
+  overrides: Partial<Record<string, unknown>> = {},
+): JSONObject {
+  return {
+    stepType: "ResourceCommand",
+    resourceId: RESOURCE_ID.toString(),
+    command: "docker restart web",
+    rationale: "the web container is wedged",
+    expectedEffect: "web serves again",
+    ...overrides,
+  } as JSONObject;
+}
+
+describe("resource round classification", () => {
+  it("a suggestion that names a resource and no rule is a resource round, never a cluster one", () => {
+    expect(
+      isResourceRemediationRound({
+        resourceType: AiResourceType.Host,
+        resourceId: RESOURCE_ID,
+      }),
+    ).toBe(true);
+    expect(
+      isClusterRemediationRound({
+        resourceType: AiResourceType.Host,
+        resourceId: RESOURCE_ID,
+      } as never),
+    ).toBe(false);
+  });
+
+  it.each([
+    [
+      "a rule round",
+      {
+        resourceType: AiResourceType.Host,
+        resourceId: RESOURCE_ID,
+        autoRemediationRuleId: ObjectID.generate(),
+      },
+    ],
+    ["a row without a type", { resourceId: RESOURCE_ID }],
+    ["a row without an id", { resourceType: AiResourceType.Host }],
+    ["a cluster round", { kubernetesClusterId: ObjectID.generate() }],
+  ])(
+    "%s is not a resource round",
+    (_label: string, row: Record<string, unknown>) => {
+      expect(isResourceRemediationRound(row as never)).toBe(false);
+    },
+  );
+});
+
+describe("RemediationExecutionRunner.checkProjectGates for a resource round", () => {
+  afterEach(() => {
+    jest.restoreAllMocks();
+  });
+
+  it("does not need the AI command execution opt-in, but still needs AI and auto-remediation", async () => {
+    const project: jest.SpyInstance = jest
+      .spyOn(ProjectService, "findOneById")
+      .mockResolvedValue({
+        enableAi: true,
+        enableAutoRemediation: true,
+        enableAiCommandExecution: false,
+      } as unknown as Project);
+
+    expect(
+      await RemediationExecutionRunner.checkProjectGates({
+        projectId: PROJECT_ID,
+        isClusterRound: false,
+        isResourceRound: true,
+      }),
+    ).toBeNull();
+
+    // A rule round in the same project is still stopped by the opt-in.
+    expect(
+      await RemediationExecutionRunner.checkProjectGates({
+        projectId: PROJECT_ID,
+        isClusterRound: false,
+      }),
+    ).toContain("AI command execution is not enabled");
+
+    project.mockResolvedValue({
+      enableAi: false,
+      enableAutoRemediation: true,
+    } as unknown as Project);
+
+    expect(
+      await RemediationExecutionRunner.checkProjectGates({
+        projectId: PROJECT_ID,
+        isClusterRound: false,
+        isResourceRound: true,
+      }),
+    ).toContain("AI or auto-remediation was disabled");
+  });
+});
+
+describe("RemediationExecutionRunner.resolveResourceMode", () => {
+  let countBy: jest.SpyInstance;
+  let jobFindBy: jest.SpyInstance;
+  let suggestionFindBy: jest.SpyInstance;
+
+  function suggestion(
+    executionMode: AutoRemediationExecutionMode,
+  ): AutoRemediationSuggestion {
+    return {
+      id: SUGGESTION_ID,
+      projectId: PROJECT_ID,
+      resourceType: AiResourceType.DockerHost,
+      resourceId: RESOURCE_ID,
+      executionMode,
+      createdAt: OneUptimeDate.getSomeMinutesAgo(1),
+    } as unknown as AutoRemediationSuggestion;
+  }
+
+  beforeEach(() => {
+    jest.spyOn(logger, "warn").mockImplementation((): void => {
+      return undefined;
+    });
+    jest.spyOn(logger, "error").mockImplementation((): void => {
+      return undefined;
+    });
+    countBy = jest
+      .spyOn(AutoRemediationSuggestionService, "countBy")
+      .mockResolvedValue(new PositiveNumber(0));
+    jobFindBy = jest.spyOn(RunnerJobService, "findBy").mockResolvedValue([]);
+    suggestionFindBy = jest
+      .spyOn(AutoRemediationSuggestionService, "findBy")
+      .mockResolvedValue([]);
+  });
+
+  afterEach(() => {
+    jest.restoreAllMocks();
+  });
+
+  it("a Suggest snapshot asks, without reading the breaker", async () => {
+    const resolution: ResourceModeResolution =
+      await RemediationExecutionRunner.resolveResourceMode({
+        suggestion: suggestion(AutoRemediationExecutionMode.Suggest),
+        resource: resource({
+          aiRemediationMode: ResourceAiRemediationMode.BypassApproval,
+        }),
+      });
+
+    expect(resolution.mode).toBe("Suggest");
+    expect(resolution.downgradedByModeChange).toBe(false);
+    expect(countBy).not.toHaveBeenCalled();
+  });
+
+  it.each([
+    ResourceAiRemediationMode.Automatic,
+    ResourceAiRemediationMode.BypassApproval,
+  ])(
+    "a FullAuto snapshot on a %s resource with headroom and no hold runs FullAuto",
+    async (mode: ResourceAiRemediationMode) => {
+      const resolution: ResourceModeResolution =
+        await RemediationExecutionRunner.resolveResourceMode({
+          suggestion: suggestion(AutoRemediationExecutionMode.FullAuto),
+          resource: resource({ aiRemediationMode: mode }),
+        });
+
+      expect(resolution).toMatchObject({
+        mode: "FullAuto",
+        downgradedByCircuitBreaker: false,
+        downgradedByModeChange: false,
+        downgradedByInFlightRound: false,
+        autoExecutedInWindow: 0,
+      });
+      // The breaker and the hold read the resource, never a cluster.
+      expect(
+        (countBy.mock.calls[0]![0] as { query: Record<string, unknown> }).query,
+      ).toMatchObject({
+        resourceType: AiResourceType.DockerHost,
+        resourceId: RESOURCE_ID,
+      });
+      expect(jobFindBy).toHaveBeenCalledTimes(2);
+    },
+  );
+
+  it("downgrades a FullAuto snapshot when the operator moved the resource to ask for approval", async () => {
+    const resolution: ResourceModeResolution =
+      await RemediationExecutionRunner.resolveResourceMode({
+        suggestion: suggestion(AutoRemediationExecutionMode.FullAuto),
+        resource: resource({
+          aiRemediationMode: ResourceAiRemediationMode.RequireApproval,
+        }),
+      });
+
+    expect(resolution.mode).toBe("Suggest");
+    expect(resolution.downgradedByModeChange).toBe(true);
+    expect(countBy).not.toHaveBeenCalled();
+  });
+
+  /*
+   * Automatic runs only a signal's FIRST round unattended. A follow-up
+   * round started FullAuto under Bypass approval must ask once the
+   * resource is on Automatic — the round is read off the round's
+   * server-written name.
+   */
+  it("downgrades a FullAuto follow-up round when the operator moved the resource from Bypass approval to Automatic", async () => {
+    const followUp: AutoRemediationSuggestion = {
+      ...suggestion(AutoRemediationExecutionMode.FullAuto),
+      ruleNameSnapshot: 'AI remediation for Docker host "web-1" (round 2)',
+    } as unknown as AutoRemediationSuggestion;
+
+    const resolution: ResourceModeResolution =
+      await RemediationExecutionRunner.resolveResourceMode({
+        suggestion: followUp,
+        resource: resource({
+          aiRemediationMode: ResourceAiRemediationMode.Automatic,
+        }),
+      });
+
+    expect(resolution.mode).toBe("Suggest");
+    expect(resolution.downgradedByModeChange).toBe(true);
+    expect(countBy).not.toHaveBeenCalled();
+  });
+
+  it("negative control: a FullAuto follow-up round still on Bypass approval runs FullAuto", async () => {
+    const followUp: AutoRemediationSuggestion = {
+      ...suggestion(AutoRemediationExecutionMode.FullAuto),
+      ruleNameSnapshot: 'AI remediation for Docker host "web-1" (round 2)',
+    } as unknown as AutoRemediationSuggestion;
+
+    const resolution: ResourceModeResolution =
+      await RemediationExecutionRunner.resolveResourceMode({
+        suggestion: followUp,
+        resource: resource({
+          aiRemediationMode: ResourceAiRemediationMode.BypassApproval,
+        }),
+      });
+
+    expect(resolution.mode).toBe("FullAuto");
+    expect(resolution.downgradedByModeChange).toBe(false);
+  });
+
+  it("negative control: a resource NAMED like a follow-up is still round 1, and Automatic runs it FullAuto", async () => {
+    const roundOne: AutoRemediationSuggestion = {
+      ...suggestion(AutoRemediationExecutionMode.FullAuto),
+      ruleNameSnapshot: 'AI remediation for Docker host "web (round 2)"',
+    } as unknown as AutoRemediationSuggestion;
+
+    const resolution: ResourceModeResolution =
+      await RemediationExecutionRunner.resolveResourceMode({
+        suggestion: roundOne,
+        resource: resource({
+          aiRemediationMode: ResourceAiRemediationMode.Automatic,
+        }),
+      });
+
+    expect(resolution.mode).toBe("FullAuto");
+  });
+
+  it("downgrades when the hourly breaker tripped, even under Bypass approval", async () => {
+    countBy.mockResolvedValue(
+      new PositiveNumber(MAX_AUTO_EXECUTIONS_PER_RULE_PER_HOUR),
+    );
+
+    const resolution: ResourceModeResolution =
+      await RemediationExecutionRunner.resolveResourceMode({
+        suggestion: suggestion(AutoRemediationExecutionMode.FullAuto),
+        resource: resource({
+          aiRemediationMode: ResourceAiRemediationMode.BypassApproval,
+        }),
+      });
+
+    expect(resolution).toMatchObject({
+      mode: "Suggest",
+      downgradedByCircuitBreaker: true,
+      breakerCheckFailed: false,
+      autoExecutedInWindow: MAX_AUTO_EXECUTIONS_PER_RULE_PER_HOUR,
+    });
+  });
+
+  it("downgrades when the breaker cannot be checked", async () => {
+    countBy.mockRejectedValue(new Error("db down"));
+
+    const resolution: ResourceModeResolution =
+      await RemediationExecutionRunner.resolveResourceMode({
+        suggestion: suggestion(AutoRemediationExecutionMode.FullAuto),
+        resource: resource(),
+      });
+
+    expect(resolution).toMatchObject({
+      mode: "Suggest",
+      downgradedByCircuitBreaker: true,
+      breakerCheckFailed: true,
+    });
+  });
+
+  it("downgrades when another round holds the resource, and names it", async () => {
+    suggestionFindBy.mockImplementation(
+      async (args: unknown): Promise<Array<AutoRemediationSuggestion>> => {
+        const query: Record<string, unknown> = (
+          args as { query: Record<string, unknown> }
+        ).query;
+        if (typeof query["status"] === "string") {
+          return [];
+        }
+        return [
+          {
+            id: OTHER_SUGGESTION_ID,
+            status: AutoRemediationSuggestionStatus.AutoExecuted,
+            verificationStatus: AutoRemediationVerificationStatus.Pending,
+            verificationDeadlineAt: OneUptimeDate.addRemoveMinutes(
+              OneUptimeDate.getCurrentDate(),
+              5,
+            ),
+            ruleNameSnapshot: 'AI remediation for Docker host "web-1"',
+          } as unknown as AutoRemediationSuggestion,
+        ];
+      },
+    );
+
+    const resolution: ResourceModeResolution =
+      await RemediationExecutionRunner.resolveResourceMode({
+        suggestion: suggestion(AutoRemediationExecutionMode.FullAuto),
+        resource: resource(),
+      });
+
+    expect(resolution.mode).toBe("Suggest");
+    expect(resolution.downgradedByInFlightRound).toBe(true);
+    expect(resolution.inFlightRound).toMatchObject({
+      suggestionId: OTHER_SUGGESTION_ID.toString(),
+      description: "applied a fix that is still being verified",
+    });
+  });
+
+  it("downgrades when the hold cannot be checked", async () => {
+    jobFindBy
+      .mockResolvedValueOnce([])
+      .mockRejectedValueOnce(new Error("db down"));
+
+    const resolution: ResourceModeResolution =
+      await RemediationExecutionRunner.resolveResourceMode({
+        suggestion: suggestion(AutoRemediationExecutionMode.FullAuto),
+        resource: resource(),
+      });
+
+    expect(resolution).toMatchObject({
+      mode: "Suggest",
+      downgradedByInFlightRound: true,
+      inFlightRound: null,
+    });
+  });
+
+  it("delegates the breaker to the rule engine's resource breaker", async () => {
+    const state: ResourceBreakerState =
+      await RemediationExecutionRunner.getResourceBreakerState({
+        resourceType: AiResourceType.Host,
+        resourceId: RESOURCE_ID.toString(),
+      });
+
+    expect(state).toEqual({ autoExecutedInWindow: 0, hasHeadroom: true });
+  });
+});
+
+describe("RemediationExecutionRunner.executeRemediation — resource rounds", () => {
+  let suggestionCas: jest.SpyInstance;
+  let suggestionUpdate: jest.SpyInstance;
+  let incidentFeed: jest.SpyInstance;
+  let statusForResource: jest.SpyInstance;
+  let countBy: jest.SpyInstance;
+  let suggestionFindBy: jest.SpyInstance;
+  let enqueue: jest.SpyInstance;
+  let project: jest.SpyInstance;
+
+  beforeEach(() => {
+    jest
+      .spyOn(AIRunService, "attemptStatusTransition")
+      .mockResolvedValue(1 as never);
+    jest
+      .spyOn(AIRunService, "updateOneBy")
+      .mockResolvedValue(undefined as never);
+    suggestionCas = jest
+      .spyOn(AutoRemediationSuggestionService, "attemptStatusTransition")
+      .mockResolvedValue(1 as never);
+    suggestionUpdate = jest
+      .spyOn(AutoRemediationSuggestionService, "updateOneById")
+      .mockResolvedValue(undefined as never);
+    incidentFeed = jest
+      .spyOn(IncidentFeedService, "createIncidentFeedItem")
+      .mockResolvedValue(undefined as never);
+    jest
+      .spyOn(AlertFeedService, "createAlertFeedItem")
+      .mockResolvedValue(undefined as never);
+    for (const level of ["debug", "warn", "error"] as const) {
+      jest.spyOn(logger, level).mockImplementation((): void => {
+        return undefined;
+      });
+    }
+    // The project did NOT opt into AI command execution: a resource round does not need it.
+    project = jest.spyOn(ProjectService, "findOneById").mockResolvedValue({
+      id: PROJECT_ID,
+      enableAi: true,
+      enableAutoRemediation: true,
+      enableAiCommandExecution: false,
+    } as unknown as Project);
+    jest.spyOn(IncidentService, "findOneById").mockResolvedValue({
+      id: INCIDENT_ID,
+      title: "web-1 keeps exiting",
+      description: "OOMKilled",
+      incidentNumber: 7,
+    } as unknown as Incident);
+    jest.spyOn(PostedRootCause, "getForSubject").mockResolvedValue(null);
+    suggestionFindBy = jest
+      .spyOn(AutoRemediationSuggestionService, "findBy")
+      .mockResolvedValue([]);
+    countBy = jest
+      .spyOn(AutoRemediationSuggestionService, "countBy")
+      .mockResolvedValue(new PositiveNumber(0));
+    jest.spyOn(RunnerJobService, "findBy").mockResolvedValue([]);
+    jest
+      .spyOn(RunnerJobService, "countBy")
+      .mockResolvedValue(new PositiveNumber(0));
+    statusForResource = jest
+      .spyOn(ResourceAiAccessService, "getStatusForResource")
+      .mockResolvedValue(resource());
+    jest
+      .spyOn(ResourceAiAccessService, "recordCommandOutcome")
+      .mockResolvedValue(undefined);
+    enqueue = jest
+      .spyOn(RunnerJobService, "enqueueAiResourceCommand")
+      .mockResolvedValue({
+        id: JOB_ID,
+        status: RunnerJobStatus.Pending,
+        payload: { displayCommand: "docker restart web", program: "docker" },
+      } as unknown as RunnerJob);
+    jest.spyOn(RunnerJobService, "pollUntilTerminal").mockResolvedValue({
+      id: JOB_ID,
+      status: RunnerJobStatus.Succeeded,
+      exitCode: 0,
+      output: "web",
+    } as unknown as RunnerJob);
+    jest
+      .spyOn(Semaphore, "lock")
+      .mockResolvedValue({ id: "lock" } as unknown as SemaphoreMutex);
+    jest.spyOn(Semaphore, "release").mockResolvedValue(undefined);
+  });
+
+  afterEach(() => {
+    jest.restoreAllMocks();
+  });
+
+  it("runs an Automatic resource's round FullAuto with the resource toolkit, the read tool, the persona and the context", async () => {
+    const findOneById: jest.SpyInstance =
+      mockSuggestionHonouringSelect(resourceRow());
+    const ruleRead: jest.SpyInstance = jest.spyOn(
+      AutoRemediationRuleService,
+      "findOneById",
+    );
+    const clusterRead: jest.SpyInstance = jest.spyOn(
+      KubernetesClusterAiAccessService,
+      "getStatusForCluster",
+    );
+    const request: { get: () => InvestigationRequest } = captureRequest();
+
+    await run();
+
+    const select: Record<string, boolean> = (
+      findOneById.mock.calls[0]![0] as { select: Record<string, boolean> }
+    ).select;
+    expect(select["resourceType"]).toBe(true);
+    expect(select["resourceId"]).toBe(true);
+
+    expect(statusForResource).toHaveBeenCalledWith({
+      projectId: PROJECT_ID,
+      resourceType: AiResourceType.DockerHost,
+      resourceId: RESOURCE_ID,
+    });
+    expect(ruleRead).not.toHaveBeenCalled();
+    expect(clusterRead).not.toHaveBeenCalled();
+
+    expect(toolNames(request.get()).sort()).toEqual(
+      [
+        "list_command_targets",
+        "execute_remediation_command",
+        LIST_INFRASTRUCTURE_ACCESS_TOOL_NAME,
+        RUN_INFRASTRUCTURE_COMMAND_TOOL_NAME,
+      ].sort(),
+    );
+    expect(request.get().personaOverride).toContain(
+      "turned on Automatic remediation",
+    );
+    expect(request.get().personaOverride).toContain('Docker host "web-1"');
+    expect(request.get().questionOverride).toContain(
+      'A signal has been declared on Docker host "web-1"',
+    );
+    expect(request.get().questionOverride).toContain("docker");
+
+    const context: string = request.get().contextSummary || "";
+    expect(context).toContain("# The resource");
+    expect(context).toContain(
+      `Docker host "web-1" (resourceId: ${RESOURCE_ID.toString()}), reached through its Docker AI agent (programs: docker)`,
+    );
+    expect(context).toContain("Remediation mode: Automatic");
+    expect(context).toContain(
+      "An unattended run becomes a proposal when the hourly per-resource circuit breaker trips",
+    );
+    expect(context).toContain("Where the Docker AI agent writes:");
+    expect(context).toContain("You may execute at most 5 commands");
+    expect(context).not.toContain("Matched auto-remediation rule");
+    expect(context).not.toContain("downgraded");
+
+    // Headroom: nothing to correct on the row, nothing on the feed.
+    expect(incidentFeed).not.toHaveBeenCalled();
+    expect(suggestionUpdate).not.toHaveBeenCalled();
+  });
+
+  it("runs a RequireApproval resource's round as a plan for approval", async () => {
+    mockSuggestionHonouringSelect(
+      resourceRow({ executionMode: AutoRemediationExecutionMode.Suggest }),
+    );
+    statusForResource.mockResolvedValue(
+      resource({
+        aiRemediationMode: ResourceAiRemediationMode.RequireApproval,
+      }),
+    );
+    const request: { get: () => InvestigationRequest } = captureRequest();
+
+    await run();
+
+    expect(toolNames(request.get())).toContain("propose_remediation_commands");
+    expect(toolNames(request.get())).not.toContain(
+      "execute_remediation_command",
+    );
+    expect(request.get().personaOverride).toContain("REMEDIATION PLANNING");
+    expect(request.get().questionOverride).toContain(
+      "compose a plan for human approval",
+    );
+    expect(request.get().contextSummary).toContain(
+      "Remediation mode: a human approves",
+    );
+  });
+
+  it("runs a BypassApproval resource with the bypass persona and question", async () => {
+    mockSuggestionHonouringSelect(resourceRow());
+    statusForResource.mockResolvedValue(
+      resource({
+        aiRemediationMode: ResourceAiRemediationMode.BypassApproval,
+      }),
+    );
+    const request: { get: () => InvestigationRequest } = captureRequest();
+
+    await run();
+
+    expect(request.get().personaOverride).toContain(
+      "chose to bypass approvals entirely",
+    );
+    expect(request.get().questionOverride).toContain(
+      "its operator bypassed approvals",
+    );
+    expect(request.get().contextSummary).toContain(
+      "Remediation mode: Bypass approval",
+    );
+  });
+
+  it("is stopped by the project kill switches like every round", async () => {
+    mockSuggestionHonouringSelect(resourceRow());
+    project.mockResolvedValue({
+      enableAi: true,
+      enableAutoRemediation: false,
+    } as unknown as Project);
+    const executeRun: jest.SpyInstance = jest.spyOn(
+      AIInvestigationEngine,
+      "executeRun",
+    );
+
+    await run();
+
+    expect(executeRun).not.toHaveBeenCalled();
+    expect(suggestionCas.mock.calls[0]![0]).toMatchObject({
+      set: { status: AutoRemediationSuggestionStatus.NoneApplicable },
+    });
+  });
+
+  it.each([
+    [
+      "no longer ready",
+      resource({
+        isRemediationReady: false,
+        gaps: [
+          {
+            code: "remediation_write_access_missing",
+            title: "The Docker AI agent is read-only",
+            nextStep: "Set ONEUPTIME_AI_ALLOW_WRITES=true.",
+            blocksInvestigation: false,
+            blocksRemediation: true,
+          },
+        ],
+      }),
+      'OneUptime AI can no longer remediate Docker host "web-1": The Docker AI agent is read-only.',
+    ],
+    [
+      "deleted",
+      null,
+      'OneUptime AI can no longer remediate Docker host "(deleted)".',
+    ],
+  ])(
+    "settles NoneApplicable, naming why, when the resource is %s",
+    async (
+      _label: string,
+      status: ResourceAiAccessStatus | null,
+      expected: string,
+    ) => {
+      mockSuggestionHonouringSelect(resourceRow());
+      statusForResource.mockResolvedValue(status);
+      const executeRun: jest.SpyInstance = jest.spyOn(
+        AIInvestigationEngine,
+        "executeRun",
+      );
+
+      await run();
+
+      expect(executeRun).not.toHaveBeenCalled();
+      const set: Record<string, unknown> = (
+        suggestionCas.mock.calls[0]![0] as { set: Record<string, unknown> }
+      ).set;
+      expect(set["status"]).toBe(
+        AutoRemediationSuggestionStatus.NoneApplicable,
+      );
+      expect(set["rationaleMarkdown"]).toContain(expected);
+      expect(set["rationaleMarkdown"]).toContain(
+        "Review the Docker host's AI agent page (AI → AI agent).",
+      );
+    },
+  );
+
+  it("a breaker downgrade corrects the row, tells the human on the feed and opens the rationale", async () => {
+    mockSuggestionHonouringSelect(resourceRow());
+    countBy.mockResolvedValue(
+      new PositiveNumber(MAX_AUTO_EXECUTIONS_PER_RULE_PER_HOUR),
+    );
+    const request: { get: () => InvestigationRequest } = captureRequest();
+
+    await run();
+
+    expect(toolNames(request.get())).toContain("propose_remediation_commands");
+    expect(suggestionUpdate).toHaveBeenCalledWith(
+      expect.objectContaining({
+        id: SUGGESTION_ID,
+        data: {
+          executionMode: AutoRemediationExecutionMode.Suggest,
+          autoResolveOnRecovery: false,
+        },
+      }),
+    );
+    const markdown: string = (
+      incidentFeed.mock.calls[0]![0] as { feedInfoInMarkdown: string }
+    ).feedInfoInMarkdown;
+    expect(markdown).toContain(
+      'AI remediation for Docker host "web-1": the hourly circuit breaker tripped',
+    );
+    expect(markdown).toContain(
+      `Docker host "web-1" already had ${MAX_AUTO_EXECUTIONS_PER_RULE_PER_HOUR} unattended AI fix(es)`,
+    );
+    expect(request.get().contextSummary).toContain(
+      'This round was downgraded to approval: The hourly circuit breaker for Docker host "web-1" tripped',
+    );
+
+    // The note opens the rationale when the round settles.
+    await request.get().postAnalysis!(postAnalysisArgs("**Summary** nothing"));
+    const set: Record<string, unknown> = (
+      suggestionCas.mock.calls[0]![0] as { set: Record<string, unknown> }
+    ).set;
+    expect(set["rationaleMarkdown"]).toMatch(
+      /^The hourly circuit breaker for Docker host "web-1" tripped/,
+    );
+  });
+
+  it("a mode change after the announcement is its own downgrade", async () => {
+    mockSuggestionHonouringSelect(resourceRow());
+    statusForResource.mockResolvedValue(
+      resource({
+        aiRemediationMode: ResourceAiRemediationMode.RequireApproval,
+      }),
+    );
+    captureRequest();
+
+    await run();
+
+    const markdown: string = (
+      incidentFeed.mock.calls[0]![0] as { feedInfoInMarkdown: string }
+    ).feedInfoInMarkdown;
+    expect(markdown).toContain(
+      'The AI remediation mode of Docker host "web-1" was changed to "Ask for approval"',
+    );
+  });
+
+  it("a follow-up round moved from Bypass approval to Automatic is downgraded, and the feed says Automatic asks for follow-ups", async () => {
+    mockSuggestionHonouringSelect(
+      resourceRow({
+        ruleNameSnapshot: 'AI remediation for Docker host "web-1" (round 2)',
+      }),
+    );
+    statusForResource.mockResolvedValue(
+      resource({ aiRemediationMode: ResourceAiRemediationMode.Automatic }),
+    );
+    captureRequest();
+
+    await run();
+
+    // The row stops claiming an unattended round.
+    expect(suggestionUpdate).toHaveBeenCalledWith(
+      expect.objectContaining({
+        data: expect.objectContaining({
+          executionMode: AutoRemediationExecutionMode.Suggest,
+          autoResolveOnRecovery: false,
+        }),
+      }),
+    );
+
+    const markdown: string = (
+      incidentFeed.mock.calls[0]![0] as { feedInfoInMarkdown: string }
+    ).feedInfoInMarkdown;
+    expect(markdown).toContain(
+      'The AI remediation mode of Docker host "web-1" was changed to "Automatic" after this follow-up round was announced as unattended',
+    );
+    expect(markdown).toContain("asks for approval of every follow-up round");
+  });
+
+  it("settles AutoExecuted with verification once the round ran a change", async () => {
+    mockSuggestionHonouringSelect(resourceRow());
+    const request: { get: () => InvestigationRequest } = captureRequest();
+
+    await run();
+
+    const outcome: ToolCallOutcome = await findTool(
+      request.get(),
+      "execute_remediation_command",
+    ).execute(resourceArgs());
+    expect(outcome.success).toBe(true);
+    expect(enqueue).toHaveBeenCalledTimes(1);
+
+    await request.get().postAnalysis!(
+      postAnalysisArgs("**Summary** restarted web"),
+    );
+
+    const cas: Record<string, unknown> = suggestionCas.mock
+      .calls[0]![0] as Record<string, unknown>;
+    expect(cas["fromStatus"]).toBe(AutoRemediationSuggestionStatus.Planning);
+    const set: Record<string, unknown> = cas["set"] as Record<string, unknown>;
+    expect(set["status"]).toBe(AutoRemediationSuggestionStatus.AutoExecuted);
+    expect(set["verificationStatus"]).toBe(
+      AutoRemediationVerificationStatus.Pending,
+    );
+    const plan: AiRemediationCommandPlan = AiRemediationCommandPlanUtil.parse(
+      set["commandPlan"] as JSONObject,
+    )!;
+    expect(plan.commands[0]).toMatchObject({
+      stepType: RunbookStepType.ResourceCommand,
+      resourceType: AiResourceType.DockerHost,
+      resourceId: RESOURCE_ID.toString(),
+      runnerId: AGENT_ID,
+    });
+    expect(plan.commands[0]!.execution?.status).toBe(
+      AiRemediationCommandExecutionStatus.Succeeded,
+    );
+
+    const markdowns: Array<string> = incidentFeed.mock.calls.map(
+      (call: Array<unknown>): string => {
+        return (call[0] as { feedInfoInMarkdown: string }).feedInfoInMarkdown;
+      },
+    );
+    expect(markdowns[markdowns.length - 1]).toContain(
+      'AI remediation for Docker host "web-1": AI executed 1 command(s).',
+    );
+    expect(markdowns.join("\n")).not.toContain("Auto Remediation Rule");
+  });
+
+  it("an unattended round that ran nothing proposes exactly the changes it kept for a human", async () => {
+    mockSuggestionHonouringSelect(resourceRow());
+    const request: { get: () => InvestigationRequest } = captureRequest();
+
+    await run();
+
+    const refused: ToolCallOutcome = await findTool(
+      request.get(),
+      "execute_remediation_command",
+    ).execute(
+      resourceArgs({
+        command: "docker stop web",
+        rollbackCommand: "docker start web",
+      }),
+    );
+    expect(refused.success).toBe(false);
+    expect(enqueue).not.toHaveBeenCalled();
+
+    await request.get().postAnalysis!(postAnalysisArgs("**Summary** stop it"));
+
+    // The row stops claiming an unattended round before it is proposed.
+    expect(suggestionUpdate).toHaveBeenCalledWith(
+      expect.objectContaining({
+        data: {
+          executionMode: AutoRemediationExecutionMode.Suggest,
+          autoResolveOnRecovery: false,
+        },
+      }),
+    );
+
+    const set: Record<string, unknown> = (
+      suggestionCas.mock.calls[0]![0] as { set: Record<string, unknown> }
+    ).set;
+    expect(set["status"]).toBe(AutoRemediationSuggestionStatus.Suggested);
+    expect(set["rationaleMarkdown"]).toContain(
+      "OneUptime AI did not run the following change(s) on its own",
+    );
+    expect(set["rationaleMarkdown"]).not.toContain("kubectl");
+    const plan: AiRemediationCommandPlan = AiRemediationCommandPlanUtil.parse(
+      set["commandPlan"] as JSONObject,
+    )!;
+    expect(plan.commands).toHaveLength(1);
+    expect(plan.commands[0]).toMatchObject({
+      sequence: 1,
+      stepType: RunbookStepType.ResourceCommand,
+      command: "docker stop web",
+      rollbackCommand: "docker start web",
+      resourceCommandTier: ResourceCommandTier.RiskyWrite,
+      policyVerdict: AiRemediationCommandPolicyVerdict.RequiresApproval,
+      runnerId: AGENT_ID,
+    });
+
+    const markdowns: Array<string> = incidentFeed.mock.calls.map(
+      (call: Array<unknown>): string => {
+        return (call[0] as { feedInfoInMarkdown: string }).feedInfoInMarkdown;
+      },
+    );
+    expect(markdowns[markdowns.length - 1]).toContain(
+      "AI needs your approval for 1 change(s) it did not run on its own",
+    );
+  });
+
+  it.each([
+    [
+      "stopped allowing remediation",
+      resource({ isRemediationReady: false }),
+      'Docker host "web-1" stopped allowing AI remediation during this round',
+    ],
+    [
+      "is reached through another agent",
+      resource({
+        agent: {
+          agentId: OTHER_AGENT_ID,
+          connectionStatus: "connected",
+          isOnline: true,
+          posture: posture(),
+        },
+      }),
+      "is no longer reached through the Docker AI agent this round composed its changes for",
+    ],
+    [
+      "has an agent that would refuse the change",
+      resource({}, { protectedTargets: ["web"] }),
+      'The Docker AI agent of Docker host "web-1" would refuse every change this round kept',
+    ],
+    ["was deleted", null, 'Docker host "web-1" was deleted during this round'],
+  ])(
+    "proposes nothing when, at settle, the resource %s",
+    async (
+      _label: string,
+      live: ResourceAiAccessStatus | null,
+      expected: string,
+    ) => {
+      mockSuggestionHonouringSelect(resourceRow());
+      const request: { get: () => InvestigationRequest } = captureRequest();
+
+      await run();
+
+      await findTool(request.get(), "execute_remediation_command").execute(
+        resourceArgs({ command: "docker stop web" }),
+      );
+
+      statusForResource.mockResolvedValue(live);
+
+      await request.get().postAnalysis!(postAnalysisArgs("**Summary** x"));
+
+      const set: Record<string, unknown> = (
+        suggestionCas.mock.calls[0]![0] as { set: Record<string, unknown> }
+      ).set;
+      expect(set["status"]).toBe(
+        AutoRemediationSuggestionStatus.NoneApplicable,
+      );
+      expect(set["rationaleMarkdown"]).toContain(expected);
+      expect(set["rationaleMarkdown"]).toContain(
+        "Review the Docker host's AI agent page (AI → AI agent).",
+      );
+    },
+  );
+
+  it("shows the model the earlier rounds on this resource for this signal", async () => {
+    mockSuggestionHonouringSelect(
+      resourceRow({
+        executionMode: AutoRemediationExecutionMode.Suggest,
+        ruleNameSnapshot: 'AI remediation for Docker host "web-1" (round 2)',
+      }),
+    );
+    suggestionFindBy.mockResolvedValue([
+      {
+        id: OTHER_SUGGESTION_ID,
+        status: AutoRemediationSuggestionStatus.AutoExecuted,
+        verificationStatus: AutoRemediationVerificationStatus.Failed,
+        verificationNote: "Monitors still down.",
+        commandPlan: {
+          commands: [
+            {
+              sequence: 1,
+              stepType: RunbookStepType.ResourceCommand,
+              runnerId: AGENT_ID,
+              runnerNameSnapshot: "Docker AI agent",
+              resourceType: AiResourceType.DockerHost,
+              resourceId: RESOURCE_ID.toString(),
+              command: "docker restart web",
+              rollbackCommand: "docker start web",
+              timeoutInMs: 60000,
+              rationale: "r",
+              expectedEffect: "e",
+              policyVerdict: AiRemediationCommandPolicyVerdict.AutoApproved,
+              execution: {
+                status: AiRemediationCommandExecutionStatus.Succeeded,
+              },
+              rollbackExecution: {
+                status: AiRemediationCommandExecutionStatus.Failed,
+              },
+            },
+          ],
+          rollbackStatus: AiRemediationRollbackStatus.Failed,
+        },
+      } as unknown as AutoRemediationSuggestion,
+    ]);
+    const request: { get: () => InvestigationRequest } = captureRequest();
+
+    await run();
+
+    const context: string = request.get().contextSummary || "";
+    expect(context).toContain(
+      "# Previous remediation attempts on this Docker host for this signal",
+    );
+    expect(context).toContain("docker restart web");
+    expect(context).toContain(
+      `WARNING: this attempt's changes may STILL BE APPLIED on the Docker host — confirm the live state with ${RUN_INFRASTRUCTURE_COMMAND_TOOL_NAME} before acting.`,
+    );
+    expect(context).toContain(
+      `diagnose the Docker host as it is now (${RUN_INFRASTRUCTURE_COMMAND_TOOL_NAME})`,
+    );
+
+    const previousQuery: Record<string, unknown> = suggestionFindBy.mock.calls
+      .map((call: Array<unknown>) => {
+        return (call[0] as { query: Record<string, unknown> }).query;
+      })
+      .find((query: Record<string, unknown>) => {
+        return "incidentId" in query;
+      }) as Record<string, unknown>;
+    expect(previousQuery).toMatchObject({
+      incidentId: INCIDENT_ID,
+      resourceType: AiResourceType.DockerHost,
+      resourceId: RESOURCE_ID,
+    });
+    expect(previousQuery["kubernetesClusterId"]).toBeUndefined();
+  });
+
+  it("the downgrade note names another round holding the resource", async () => {
+    mockSuggestionHonouringSelect(resourceRow());
+    suggestionFindBy.mockImplementation(
+      async (args: unknown): Promise<Array<AutoRemediationSuggestion>> => {
+        const query: Record<string, unknown> = (
+          args as { query: Record<string, unknown> }
+        ).query;
+        if (typeof query["status"] === "string" || "incidentId" in query) {
+          return [];
+        }
+        return [
+          {
+            id: OTHER_SUGGESTION_ID,
+            status: AutoRemediationSuggestionStatus.Planning,
+            executionMode: AutoRemediationExecutionMode.FullAuto,
+            createdAt: OneUptimeDate.getSomeMinutesAgo(30),
+            ruleNameSnapshot: 'AI remediation for Docker host "web-1"',
+          } as unknown as AutoRemediationSuggestion,
+        ];
+      },
+    );
+    const request: { get: () => InvestigationRequest } = captureRequest();
+
+    await run();
+
+    expect(request.get().contextSummary).toContain(
+      'another OneUptime AI round on Docker host "web-1" (AI remediation for Docker host "web-1") is still running unattended',
+    );
+    expect(request.get().contextSummary).toContain(
+      "This round was downgraded to approval",
+    );
+    // A Suggest round's context no longer promises an unattended run.
+    expect(request.get().contextSummary).not.toContain(
+      RESOURCE_UNATTENDED_ROUND_BECOMES_PROPOSAL_SUMMARY,
+    );
+  });
+});