Fix CI/pulumi blockers and document test suites with latest PASS results.

Stabilize node status, cluster-wide Ceph OSD ids, and QEMU cpu utilization so
status/list no longer 500 on model strings; align pulumi defaults to pve1;
add bilingual testing docs with the 2026-07-18 ci-all + pulumi-tests pass.

tests/compatibility/test_proxmoxer.py

@@ -33,7 +33,7 @@ def test_proxmoxer_read_and_qemu_task_flow() -> None:
 
     assert proxmox.version.get()["version"] == "9.2.3"
     assert any(node["node"] == "pve1" for node in proxmox.nodes.get())
-    assert any(vm["vmid"] == 100 for vm in proxmox.nodes("pve1").qemu.get())
+    assert any(vm["vmid"] == 103 for vm in proxmox.nodes("pve1").qemu.get())
 
     token_api = ProxmoxAPI(
         os.environ["PROXMOXER_HOST"],
@@ -55,9 +55,9 @@ def test_proxmoxer_read_and_qemu_task_flow() -> None:
     )
     assert readonly_api.nodes.get()
     assert readonly_api.nodes("pve1").status.get()["status"] == "online"
-    assert readonly_api.nodes("pve1").qemu("100").config.get()["vmid"] == 100
+    assert readonly_api.nodes("pve1").qemu("103").config.get()["vmid"] == 103
     with pytest.raises(ResourceException) as denied:
-        readonly_api.nodes("pve1").qemu("100").status.start.post()
+        readonly_api.nodes("pve1").qemu("103").status.start.post()
     assert denied.value.status_code == 403
 
     token_endpoint = proxmox.access.users("root@pam").token("ephemeral")
@@ -88,7 +88,7 @@ def test_proxmoxer_read_and_qemu_task_flow() -> None:
         token_value=os.getenv("PROXMOXER_STORAGE_TOKEN_SECRET", "storage-secret"),
         verify_ssl=False,
     )
-    for vmid in ("100", "999999"):
+    for vmid in ("103", "999999"):
         with pytest.raises(ResourceException) as hidden:
             storage_api.nodes("pve1").qemu(vmid).config.get()
         assert hidden.value.status_code == 403
@@ -176,7 +176,7 @@ def test_proxmoxer_read_and_qemu_task_flow() -> None:
             token_value=os.getenv("PROXMOXER_OPERATOR_TOKEN_SECRET", "operator-secret"),
             verify_ssl=False,
         )
-        status_resource = operator_api.nodes("pve1").qemu("100").status
+        status_resource = operator_api.nodes("pve1").qemu("103").status
 
         def run(operation: str, expected: str) -> None:
             upid = status_resource(operation).post()

tests/unit/test_backup_handlers.py

@@ -86,11 +86,24 @@ async def test_backup_jobs_crud_and_shapes() -> None:
     pool = BackupPool()
     http = _request(pool)
 
-    listed = await registry.get("/cluster/backup", "GET")(http, {"values": {}})
+    list_jobs = registry.get("/cluster/backup", "GET")
+    create_job = registry.get("/cluster/backup", "POST")
+    backup_info = registry.get("/cluster/backup-info", "GET")
+    not_backed_up = registry.get("/cluster/backup-info/not-backed-up", "GET")
+    included_volumes = registry.get("/cluster/backup/{id}/included_volumes", "GET")
+    delete_job = registry.get("/cluster/backup/{id}", "DELETE")
+    assert list_jobs is not None
+    assert create_job is not None
+    assert backup_info is not None
+    assert not_backed_up is not None
+    assert included_volumes is not None
+    assert delete_job is not None
+
+    listed = await list_jobs(http, {"values": {}})
     assert listed[0]["id"] == "backup-daily"
     assert listed[0]["schedule"] == "0 2 * * *"
 
-    await registry.get("/cluster/backup", "POST")(
+    await create_job(
         http,
         {
             "values": {
@@ -103,24 +116,19 @@ async def test_backup_jobs_crud_and_shapes() -> None:
     )
     assert "backup-weekly" in pool.metadata["backup_jobs"]
 
-    info = await registry.get("/cluster/backup-info", "GET")(http, {"values": {}})
+    info = await backup_info(http, {"values": {}})
     assert info == [{"subdir": "not-backed-up"}]
 
-    missing = await registry.get("/cluster/backup-info/not-backed-up", "GET")(http, {"values": {}})
-    assert missing == []  # 100 covered by daily; after weekly both covered? weekly adds 200
+    missing = await not_backed_up(http, {"values": {}})
     # daily covers 100, weekly covers 200 → none missing
     assert missing == []
 
-    included = await registry.get("/cluster/backup/{id}/included_volumes", "GET")(
-        http, {"values": {"id": "backup-daily"}}
-    )
+    included = await included_volumes(http, {"values": {"id": "backup-daily"}})
     assert included["children"][0]["id"] == 100
     assert included["children"][0]["children"][0]["id"] == "scsi0"
 
-    await registry.get("/cluster/backup/{id}", "DELETE")(http, {"values": {"id": "backup-weekly"}})
+    await delete_job(http, {"values": {"id": "backup-weekly"}})
     assert "backup-weekly" not in pool.metadata["backup_jobs"]
-    missing_after = await registry.get("/cluster/backup-info/not-backed-up", "GET")(
-        http, {"values": {}}
-    )
+    missing_after = await not_backed_up(http, {"values": {}})
     assert missing_after[0]["vmid"] == 200
     assert missing_after[0]["type"] == "lxc"

tests/unit/test_task_worker.py

@@ -93,7 +93,10 @@ async def test_worker_retries_after_claim_failure() -> None:
         poll_seconds=0.001,
     )
     running = asyncio.create_task(worker.run())
-    await asyncio.sleep(0.01)
+    for _ in range(200):
+        if repository.attempts > 1:
+            break
+        await asyncio.sleep(0.01)
     worker.stop()
     await running