fix(svc-ai-ollama): preload a model that stops generating

The matrix variant preloaded smollm2:135m, which is too small to emit a
stop token reliably. Ollama's own log for the prompt "hi" shows it running
away rather than running slowly:

  n_gen = 22011, tg = 86.14 t/s
  slot context shift, n_keep = 4, n_left = 4091, n_discard = 2045
  aborting completion request due to client closing the connection

86 tokens per second is fast; 22000 tokens with repeated context shifts is
a model that never finishes. Every completion without a hard token cap ran
until the consumer's own timeout, which is why zammad reported
AI::Provider::ResponseError after exactly 300s while short warm-ups and
provider pings returned 200 immediately.

Preload qwen2.5:0.5b instead: a comparable footprint (397MB vs 270MB) that
is instruction-tuned and terminates.

web-app-openwebui asserted the served model with a hardcoded
id.startsWith("smollm2"), so it would have failed on any model change.
Render the deployed preload model into its Playwright env and compare
against that instead of a literal family name.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>

roles/svc-ai-ollama/meta/variants.yml

@@ -12,6 +12,6 @@
       enabled: true
     ollama:
       preload_models:
-        - smollm2:135m
+        - qwen2.5:0.5b
       mem_reservation: 1g
       mem_limit: 2g

roles/web-app-openwebui/files/playwright/test-ollama-chat.js

@@ -32,10 +32,11 @@ exports.register = function (shared) {
     const modelsBody = await modelsResp.json();
     const modelIds = (Array.isArray(modelsBody?.data) ? modelsBody.data : [])
       .map((m) => String(m?.id ?? m?.name ?? ""));
-    const ollamaModel = modelIds.find((id) => id.startsWith("smollm2"));
+    const preloadedModel = String(process.env.OLLAMA_PRELOAD_MODEL || "").replace(/^"|"$/g, "");
+    const ollamaModel = modelIds.find((id) => id === preloadedModel || id.startsWith(preloadedModel.split(":")[0]));
     expect(
       ollamaModel,
-      `the preloaded Ollama model (smollm2) must be served by OpenWebUI, proving the Ollama backend is reachable and the model pulled (got ${JSON.stringify(modelIds)})`
+      `the preloaded Ollama model (${preloadedModel}) must be served by OpenWebUI, proving the Ollama backend is reachable and the model pulled (got ${JSON.stringify(modelIds)})`
     ).toBeTruthy();
 
     const chatResp = await page.request.post(`${base}/api/chat/completions`, {

roles/web-app-openwebui/templates/playwright.env.j2

@@ -44,6 +44,8 @@ MCP_SERVICE_ENABLED={{ lookup('config', application_id, 'mcp.enabled') | string
 
 OIDC_ISSUER_URL={{ OIDC.CLIENT.ISSUER_URL | dotenv_quote }}
 
+OLLAMA_PRELOAD_MODEL={{ (lookup('config', 'svc-ai-ollama', 'services.ollama.preload_models') | reject('search', 'embed') | list | first) | dotenv_quote }}
+
 OPENWEBUI_BASE_URL={{ lookup('tls', application_id, 'url.base') | dotenv_quote }}
 
 SEAWEEDFS_ADMIN_PASSWORD={{ lookup('users', 'administrator').password | dotenv_quote }}