diff --git a/backend/open_webui/main.py b/backend/open_webui/main.py
index c84a394b87..7d47d12eda 100644
--- a/backend/open_webui/main.py
+++ b/backend/open_webui/main.py
@@ -417,8 +417,10 @@ async def lifespan(app: FastAPI):
         except Exception as e:
             log.warning(f'Failed to pre-fetch models at startup: {e}')
 
-    # Pre-fetch tool server specs so the first request doesn't pay the latency cost
-    if len(await Config.get('tool_server.connections', []) or []) > 0:
+    # Pre-fetch tool/terminal server specs so the first request doesn't pay the latency cost
+    has_tool_servers = len(await Config.get('tool_server.connections', []) or []) > 0
+    has_terminal_servers = len(await Config.get('terminal_server.connections', []) or []) > 0
+    if has_tool_servers or has_terminal_servers:
         mock_request = Request(
             {
                 'type': 'http',
@@ -435,18 +437,20 @@ async def lifespan(app: FastAPI):
             }
         )
 
-        log.info('Initializing tool servers...')
-        try:
-            await set_tool_servers(mock_request)
-            log.info('Initialized %s tool server(s)', len(app.state.TOOL_SERVERS))
-        except Exception as e:
-            log.warning(f'Failed to initialize tool servers at startup: {e}')
+        if has_tool_servers:
+            log.info('Initializing tool servers...')
+            try:
+                await set_tool_servers(mock_request)
+                log.info('Initialized %s tool server(s)', len(app.state.TOOL_SERVERS))
+            except Exception as e:
+                log.warning(f'Failed to initialize tool servers at startup: {e}')
 
-        try:
-            await set_terminal_servers(mock_request)
-            log.info('Initialized %s terminal server(s)', len(app.state.TERMINAL_SERVERS))
-        except Exception as e:
-            log.warning(f'Failed to initialize terminal servers at startup: {e}')
+        if has_terminal_servers:
+            try:
+                await set_terminal_servers(mock_request)
+                log.info('Initialized %s terminal server(s)', len(app.state.TERMINAL_SERVERS))
+            except Exception as e:
+                log.warning(f'Failed to initialize terminal servers at startup: {e}')
 
     # Mark application as ready to accept traffic from a startup perspective.
     if license_task:
diff --git a/backend/open_webui/utils/tools.py b/backend/open_webui/utils/tools.py
index 4ab5c96504..8ed6c4ae67 100644
--- a/backend/open_webui/utils/tools.py
+++ b/backend/open_webui/utils/tools.py
@@ -1299,7 +1299,36 @@ async def set_terminal_servers(request: Request):
             }
         )
 
-    request.app.state.TERMINAL_SERVERS = await get_tool_servers_data(server_configs)
+    terminal_servers = await get_tool_servers_data(server_configs)
+
+    # A failed spec fetch (e.g. cold orchestrator right after a restart) must
+    # not evict a previously cached spec: keep the last known good entry for
+    # any enabled connection that dropped out instead of caching its absence.
+    expected_urls = {
+        server_config['info']['id'] or str(idx): server_config['url']
+        for idx, server_config in enumerate(server_configs)
+        if server_config['config']['enable']
+    }
+    missing_ids = set(expected_urls) - {server.get('id') for server in terminal_servers}
+    if missing_ids:
+        previous_servers = getattr(request.app.state, 'TERMINAL_SERVERS', None) or []
+        if request.app.state.redis is not None:
+            try:
+                data = await request.app.state.redis.get(f'{REDIS_KEY_PREFIX}:terminal_servers')
+                if data is not None:
+                    previous_servers = JSONCodec.loads(data)
+            except Exception as e:
+                log.error(f'Error fetching terminal_servers from Redis: {e}')
+        for server in previous_servers:
+            server_id = server.get('id')
+            if server_id in missing_ids and server.get('url') == expected_urls[server_id]:
+                log.warning(
+                    "Spec fetch for terminal server '%s' failed; keeping previously cached specs",
+                    server_id,
+                )
+                terminal_servers.append(server)
+
+    request.app.state.TERMINAL_SERVERS = terminal_servers
 
     # Fetch system prompts concurrently (runs at cache time, not per-request)
     connections_by_id = {c.get('id'): c for c in connections if c.get('id')}
@@ -1323,9 +1352,12 @@ async def set_terminal_servers(request: Request):
     )
 
     if request.app.state.redis is not None:
-        await request.app.state.redis.set(
-            f'{REDIS_KEY_PREFIX}:terminal_servers', JSONCodec.dumps(request.app.state.TERMINAL_SERVERS)
-        )
+        try:
+            await request.app.state.redis.set(
+                f'{REDIS_KEY_PREFIX}:terminal_servers', JSONCodec.dumps(request.app.state.TERMINAL_SERVERS)
+            )
+        except Exception as e:
+            log.error(f'Error caching terminal_servers to Redis: {e}')
 
     return request.app.state.TERMINAL_SERVERS
 
@@ -1375,6 +1407,11 @@ async def get_terminal_tools(
     # Find the cached spec data for this terminal
     terminal_servers = await get_terminal_servers(request)
     server_data = next((s for s in terminal_servers if s.get('id') == terminal_id), None)
+    if server_data is None:
+        # The cache can lack this server when its spec fetch failed earlier
+        # (e.g. it was still starting up); rebuild once before giving up.
+        terminal_servers = await set_terminal_servers(request)
+        server_data = next((s for s in terminal_servers if s.get('id') == terminal_id), None)
     if server_data is None:
         raise RuntimeError(f"Terminal server '{terminal_id}' is unavailable")