@halcyonlabs/app 0.47.0-halcyon.101 → 0.47.0-halcyon.106

package/dist/index.d.ts

@@ -155,10 +155,10 @@ declare const uiPreferenceDefinitions: {
     }>>;
     readonly "sidebar.threadGrouping.environment": UiPreferenceDefinition<z$1.ZodUnion<readonly [z$1.ZodLiteral<"auto">, z$1.ZodBoolean]>>;
     readonly "sidebar.chronologicalSort": UiPreferenceDefinition<z$1.ZodEnum<{
+        none: "none";
         updated: "updated";
         created: "created";
         alpha: "alpha";
-        none: "none";
     }>>;
     readonly "sidebar.sortDirection": UiPreferenceDefinition<z$1.ZodEnum<{
         default: "default";
@@ -451,8 +451,8 @@ type ChangedMessage = z$1.infer<typeof changedMessageSchema>;
 
 declare const environmentStatusSchema: z$1.ZodEnum<{
     error: "error";
-    provisioning: "provisioning";
     ready: "ready";
+    provisioning: "provisioning";
     creating: "creating";
     destroyed: "destroyed";
 }>;
@@ -471,8 +471,8 @@ declare const environmentSchema: z$1.ZodObject<{
     mergeBaseBranch: z$1.ZodNullable<z$1.ZodString>;
     status: z$1.ZodEnum<{
         error: "error";
-        provisioning: "provisioning";
         ready: "ready";
+        provisioning: "provisioning";
         creating: "creating";
         destroyed: "destroyed";
     }>;
@@ -4282,8 +4282,8 @@ declare const threadTimelinePendingTodosSchema: z$1.ZodObject<{
 type ThreadTimelinePendingTodos = z$1.infer<typeof threadTimelinePendingTodosSchema>;
 
 declare const threadParentNotificationsSchema: z$1.ZodEnum<{
-    auto: "auto";
     off: "off";
+    auto: "auto";
     stuck: "stuck";
 }>;
 type ThreadParentNotifications = z$1.infer<typeof threadParentNotificationsSchema>;
@@ -7456,8 +7456,8 @@ declare const hostDaemonCommandRegistry: {
             skillFilePath: z$1.ZodString;
         }, z$1.core.$strict>], "kind">>;
         instructionMode: z$1.ZodEnum<{
-            replace: "replace";
             append: "append";
+            replace: "replace";
         }>;
         type: z$1.ZodLiteral<"thread.rewind.prepare">;
         leaseId: z$1.ZodString;
@@ -7605,8 +7605,8 @@ declare const hostDaemonCommandRegistry: {
             skillFilePath: z$1.ZodString;
         }, z$1.core.$strict>], "kind">>;
         instructionMode: z$1.ZodEnum<{
-            replace: "replace";
             append: "append";
+            replace: "replace";
         }>;
         type: z$1.ZodLiteral<"thread.start">;
         requestId: z$1.ZodString;
@@ -8069,8 +8069,8 @@ declare const hostDaemonCommandRegistry: {
                 envPassthrough: z$1.ZodArray<z$1.ZodString>;
             }, z$1.core.$strict>;
             instructionMode: z$1.ZodEnum<{
-                replace: "replace";
                 append: "append";
+                replace: "replace";
             }>;
             projectId: z$1.ZodString;
             providerId: z$1.ZodString;
@@ -8285,8 +8285,8 @@ declare const hostDaemonCommandRegistry: {
                 envPassthrough: z$1.ZodArray<z$1.ZodString>;
             }, z$1.core.$strict>;
             instructionMode: z$1.ZodEnum<{
-                replace: "replace";
                 append: "append";
+                replace: "replace";
             }>;
             projectId: z$1.ZodString;
             providerId: z$1.ZodString;
@@ -9192,9 +9192,9 @@ declare const hostDaemonCommandRegistry: {
             base64: "base64";
             utf8: "utf8";
         }>;
-        sha256: z$1.ZodString;
-        sizeBytes: z$1.ZodNumber;
         mimeType: z$1.ZodOptional<z$1.ZodString>;
+        sizeBytes: z$1.ZodNumber;
+        sha256: z$1.ZodString;
         modifiedAtMs: z$1.ZodOptional<z$1.ZodNumber>;
         notModified: z$1.ZodLiteral<true>;
     }, z$1.core.$strip>]>, "onlineRpc", true>;
@@ -9459,10 +9459,10 @@ declare const hostDaemonCommandRegistry: {
         health: z$1.ZodObject<{
             status: z$1.ZodEnum<{
                 unknown: "unknown";
-                ready: "ready";
                 not_installed: "not_installed";
                 unauthenticated: "unauthenticated";
                 expired: "expired";
+                ready: "ready";
                 unsupported_version: "unsupported_version";
             }>;
             statusMessage: z$1.ZodNullable<z$1.ZodString>;
@@ -9510,9 +9510,9 @@ declare const hostDaemonCommandRegistry: {
         executablePath: z$1.ZodNullable<z$1.ZodString>;
         installed: z$1.ZodBoolean;
         installSource: z$1.ZodEnum<{
+            external: "external";
             notInstalled: "notInstalled";
             npmGlobal: "npmGlobal";
-            external: "external";
         }>;
         currentVersion: z$1.ZodNullable<z$1.ZodString>;
         latestVersion: z$1.ZodNullable<z$1.ZodString>;
@@ -9958,8 +9958,8 @@ declare const hostDaemonCommandRegistry: {
         message: z$1.ZodNullable<z$1.ZodString>;
         state: z$1.ZodNullable<z$1.ZodEnum<{
             pending: "pending";
-            activating: "activating";
             ready: "ready";
+            activating: "activating";
         }>>;
     }, z$1.core.$strict>, "onlineRpc", true>;
     "server_move.prepare": HostDaemonCommandDescriptor<"server_move.prepare", z$1.ZodObject<{
@@ -10061,9 +10061,9 @@ declare const providerCliStatusResponseSchema: z$1.ZodRecord<z$1.ZodString, z$1.
     executablePath: z$1.ZodNullable<z$1.ZodString>;
     installed: z$1.ZodBoolean;
     installSource: z$1.ZodEnum<{
+        external: "external";
         notInstalled: "notInstalled";
         npmGlobal: "npmGlobal";
-        external: "external";
     }>;
     currentVersion: z$1.ZodNullable<z$1.ZodString>;
     latestVersion: z$1.ZodNullable<z$1.ZodString>;

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@halcyonlabs/app",
-  "version": "0.47.0-halcyon.101",
+  "version": "0.47.0-halcyon.106",
   "description": "bb app launcher for npx",
   "keywords": [
     "bb",

package/server/dist/builtin-plugins/orbit/dist/assets/orbit/bundle/scripts/orbit-job-supervisor

@@ -22,7 +22,10 @@ START_HEADROOM = 5_000_000_000
 # to free file entries on filesystems that have a fixed number of them.
 INODE_FLOOR = 500_000
 START_INODE_HEADROOM = 500_000
-CONCURRENT_JOBS = 2
+MIN_CONCURRENT_JOBS = 2
+MAX_CONCURRENT_JOBS = 8
+JOB_MEMORY_BYTES = 6 * 2**30
+DESKTOP_MEMORY_BYTES = 6 * 2**30
 CAPACITY_FULL_STATUS = 75
 LIMITS = dict(retained_bytes=RETAINED_JOB_BYTES)
 # Records written before the free-space rule carry the old fixed caps. They load
@@ -213,6 +216,61 @@ def write_record(directory, name, value, exclusive=False):
         raise
 
 
+def machine_concurrency(memory, cores, override):
+    if override is not None:
+        if not re.fullmatch('[0-9]{1,9}', override):
+            raise ValueError('Job concurrency override must be a positive whole number.')
+        requested = int(override)
+        if requested < 1:
+            raise ValueError('Job concurrency override must be a positive whole number.')
+        slots = max(MIN_CONCURRENT_JOBS, min(MAX_CONCURRENT_JOBS, requested))
+        return slots, f'override {requested}; floor {MIN_CONCURRENT_JOBS}, ceiling {MAX_CONCURRENT_JOBS}'
+    memory_slots = max(0, memory - DESKTOP_MEMORY_BYTES) // JOB_MEMORY_BYTES
+    core_slots = cores // 3
+    slots = max(MIN_CONCURRENT_JOBS, min(MAX_CONCURRENT_JOBS, memory_slots, core_slots))
+    return slots, (f'{memory / 2**30:.1f} GiB RAM minus 6 GiB for desktop, antfly and main session; '
+                   f'6 GiB per job gives {memory_slots} slots; {cores} cores / 3 gives {core_slots}; '
+                   f'floor {MIN_CONCURRENT_JOBS}, ceiling {MAX_CONCURRENT_JOBS}')
+
+
+def machine_resources():
+    if sys.platform == 'darwin':
+        import subprocess
+        try:
+            memory = int(subprocess.check_output(['sysctl', '-n', 'hw.memsize'], timeout=5))
+        except subprocess.SubprocessError as error:
+            raise OSError('Machine memory reading failed.') from error
+    else:
+        memory = os.sysconf('SC_PHYS_PAGES') * os.sysconf('SC_PAGE_SIZE')
+    cores = os.cpu_count()
+    if memory <= 0 or cores is None or cores <= 0:
+        raise ValueError('Machine memory or core count is unavailable.')
+    return memory, cores
+
+
+def concurrent_jobs(state):
+    try:
+        fd = os.open('concurrency', os.O_RDONLY | os.O_NOFOLLOW | os.O_CLOEXEC | os.O_NONBLOCK, dir_fd=state)
+    except FileNotFoundError:
+        override = None
+    else:
+        with os.fdopen(fd) as stream:
+            metadata = os.fstat(stream.fileno())
+            if not stat.S_ISREG(metadata.st_mode) or metadata.st_size > 32:
+                raise ValueError('Invalid job concurrency override file.')
+            override = stream.read(33).strip()
+    try:
+        memory, cores = machine_resources()
+        slots, reason = machine_concurrency(memory, cores, override)
+    except (OSError, ValueError) as error:
+        if override is not None:
+            slots, reason = machine_concurrency(0, 0, override)
+        else:
+            slots, reason = MIN_CONCURRENT_JOBS, f'machine reading unavailable: {error}; floor {MIN_CONCURRENT_JOBS}'
+    print(f'Job concurrency: {slots}; {reason}.', file=sys.stderr)
+    return slots
+
+
 class Workspaces:
     """A single supervisor owns these roots, inaccessible to workers.
 
@@ -279,7 +337,7 @@ class Workspaces:
                     self.environment(existing['id'])
                 return existing
             slots = [r for r in records if holds_slot(r)]
-            if len(slots) >= CONCURRENT_JOBS:
+            if len(slots) >= concurrent_jobs(self.state):
                 raise CapacityFull('Job capacity is reserved; wait for disposal.')
             if not workspace:
                 if type(free) is not int or free < MAIN_FREE_FLOOR:
@@ -528,7 +586,7 @@ class Workspaces:
                 return manifest
             if manifest['state'] != 'parked':
                 raise ValueError('Only a parked job can take a slot again.')
-            if len([r for r in self.records() if holds_slot(r)]) >= CONCURRENT_JOBS:
+            if len([r for r in self.records() if holds_slot(r)]) >= concurrent_jobs(self.state):
                 raise CapacityFull('Job capacity is reserved; wait for disposal.')
             return self.save_transition(manifest, 'running')
 
@@ -1267,6 +1325,10 @@ def managed_operation(store, account, request):
         finally:
             os.close(root)
     elif request['operation'] == 'measure':
+        try:
+            concurrent_jobs(store.state)
+        except (OSError, ValueError) as error:
+            print(f'Job concurrency setting could not be read: {error}', file=sys.stderr)
         if request.get('size', True) not in (True, False):
             raise ValueError('Invalid measure request.')
         print(json.dumps(store.measure(request['id'], sized=request.get('size', True))))

package/server/dist/builtin-plugins/orbit/dist/assets/orbit/recipes/base-linux/antfly.sh

@@ -57,9 +57,9 @@ After=network.target
 User=vectors
 Group=vectors
 Environment=HOME=/var/db/orbit/vectors
-# The embedding model is loaded at start (about two minutes on two vCPUs): loaded on
-# demand, the first index creation's probe times out while it loads.
-ExecStart=/usr/local/bin/antfly standalone --host 127.0.0.1 --port 8080 --health-port 4200 --data-dir $data --models-dir $models --preload-model embedder:$model
+ExecStart=/usr/local/bin/antfly standalone --host 127.0.0.1 --port 8080 --health-port 4200 --data-dir $data --models-dir $models --process-memory-budget-mb 6144 --inference-kv-budget-mb 64 --preload-model embedder:$model
+MemoryAccounting=true
+MemoryMax=6144M
 Restart=always
 RestartSec=3
 NoNewPrivileges=true

package/server/dist/builtin-plugins/orbit/dist/assets/orbit/revision

@@ -1 +1 @@
-f1059d04b641d9927b443da4711ba9c7639a9d7b
+204aea94877ae0fdc63ccee1e6e19e67f309dc03