fix(heartbeat): prevent false process_lost failures on queued and non-child-process runs

- reapOrphanedRuns() now only scans running runs; queued runs are
  legitimately absent from runningProcesses (waiting on concurrency
  limits or issue locks) so including them caused false process_lost
  failures (closes #90)
- Add module-level activeRunExecutions set so non-child-process adapters
  (http, openclaw) are protected from the reaper during execution
- Add resumeQueuedRuns() to restart persisted queued runs after a server
  restart, called at startup and each periodic tick
- Add outer catch in executeRun() so setup failures (ensureRuntimeState,
  resolveWorkspaceForRun, etc.) are recorded as failed runs instead of
  leaving them stuck in running state
- Guard resumeQueuedRuns() against paused/terminated/pending_approval agents
- Increase opencode models discovery timeout from 20s to 45s

server/src/index.ts

@@ -513,11 +513,14 @@ export async function startServer(): Promise<StartedServer> {
   if (config.heartbeatSchedulerEnabled) {
     const heartbeat = heartbeatService(db as any);
   
-    // Reap orphaned runs at startup (no threshold -- runningProcesses is empty)
-    void heartbeat.reapOrphanedRuns().catch((err) => {
-      logger.error({ err }, "startup reap of orphaned heartbeat runs failed");
-    });
-
+    // Reap orphaned running runs at startup while in-memory execution state is empty,
+    // then resume any persisted queued runs that were waiting on the previous process.
+    void heartbeat
+      .reapOrphanedRuns()
+      .then(() => heartbeat.resumeQueuedRuns())
+      .catch((err) => {
+        logger.error({ err }, "startup heartbeat recovery failed");
+      });
     setInterval(() => {
       void heartbeat
         .tickTimers(new Date())
@@ -530,11 +533,13 @@ export async function startServer(): Promise<StartedServer> {
           logger.error({ err }, "heartbeat timer tick failed");
         });
   
-      // Periodically reap orphaned runs (5-min staleness threshold)
+      // Periodically reap orphaned runs (5-min staleness threshold) and make sure
+      // persisted queued work is still being driven forward.
       void heartbeat
         .reapOrphanedRuns({ staleThresholdMs: 5 * 60 * 1000 })
+        .then(() => heartbeat.resumeQueuedRuns())
         .catch((err) => {
-          logger.error({ err }, "periodic reap of orphaned heartbeat runs failed");
+          logger.error({ err }, "periodic heartbeat recovery failed");
         });
     }, config.heartbeatSchedulerIntervalMs);
   }

server/src/services/heartbeat.ts

@@ -1089,6 +1089,9 @@ export function heartbeatService(db: Db) {
       run = claimed;
     }
 
+    activeRunExecutions.add(run.id);
+
+    try {
     const agent = await getAgent(run.agentId);
     if (!agent) {
       await setRunStatus(runId, "failed", {
@@ -1676,10 +1679,39 @@ export function heartbeatService(db: Db) {
       }
 
       await finalizeAgentStatus(agent.id, "failed");
-    } finally {
-      await releaseRuntimeServicesForRun(run.id);
-      await startNextQueuedRunForAgent(agent.id);
-    }
+        } catch (outerErr) {
+          // Setup code before adapter.execute threw (e.g. ensureRuntimeState, resolveWorkspaceForRun).
+          // The inner catch did not fire, so we must record the failure here.
+          const message = outerErr instanceof Error ? outerErr.message : "Unknown setup failure";
+          logger.error({ err: outerErr, runId }, "heartbeat execution setup failed");
+          await setRunStatus(runId, "failed", {
+            error: message,
+            errorCode: "adapter_failed",
+            finishedAt: new Date(),
+          }).catch(() => undefined);
+          await setWakeupStatus(run.wakeupRequestId, "failed", {
+            finishedAt: new Date(),
+            error: message,
+          }).catch(() => undefined);
+          const failedRun = await getRun(runId).catch(() => null);
+          if (failedRun) {
+            // Emit a run-log event so the failure is visible in the run timeline,
+            // consistent with what the inner catch block does for adapter failures.
+            await appendRunEvent(failedRun, 1, {
+              eventType: "error",
+              stream: "system",
+              level: "error",
+              message,
+            }).catch(() => undefined);
+            await releaseIssueExecutionAndPromote(failedRun).catch(() => undefined);
+          }
+          // Ensure the agent is not left stuck in "running" if the inner catch handler's
+          // DB calls threw (e.g. a transient DB error in finalizeAgentStatus).
+          await finalizeAgentStatus(run.agentId, "failed").catch(() => undefined);
+        } finally {
+          await releaseRuntimeServicesForRun(run.id).catch(() => undefined);
+      activeRunExecutions.delete(run.id);
+          await startNextQueuedRunForAgent(run.agentId);
   }
 
   async function releaseIssueExecutionAndPromote(run: typeof heartbeatRuns.$inferSelect) {