diff --git a/CONTRIBUTING.md b/CONTRIBUTING.md
index f0ba140..49176b2 100644
--- a/CONTRIBUTING.md
+++ b/CONTRIBUTING.md
@@ -57,6 +57,8 @@ Log a short fixed message and put the values in `extra`: `log.info("poster uploa
 | `build_fonts.py` | Fetches the fallback fonts for scripts Inter lacks. | Internet |
 | `build_networks.py` | Rebuilds `assets/networks.json` and `assets/provider_networks.json` from TMDB. | `TMDB_API_KEY` |
 | `ocr_readings.py` | Adds real OCR readings to `tests/ocr_readings.json`. | `TMDB_API_KEY` |
+| `unraid_template.py` | Writes the settings help and `<Changes>` in `templates/posteryard.xml` from `README.md` and `CHANGELOG.md`. `--check` exits 1 when the settings are out of date. | |
+| `dockerhub_description.py` | Prints `README.md` as the Docker Hub overview. | |
 
 Run each with `uv run python tools/<script>`.
 
diff --git a/src/posteryard/plex.py b/src/posteryard/plex.py
index 2356e17..57b01b0 100644
--- a/src/posteryard/plex.py
+++ b/src/posteryard/plex.py
@@ -300,7 +300,14 @@ def _guide_walk(self, provider: str) -> list[set[str]]:
         batches: list[set[str]] = []
         while len(batches) < GUIDE_ID_BATCHES:
             start = 1 + len(batches) * GUIDE_IDS
-            guids = self._guide_guids(provider, start, start + GUIDE_IDS)
+            try:
+                guids = self._guide_guids(provider, start, start + GUIDE_IDS)
+            except http.RequestError as exc:
+                log.warning(
+                    "the guide ids could not all be read, some titles may get no poster",
+                    extra={"ids": start - 1, "reason": str(exc)},
+                )
+                return batches
             if not guids:
                 return batches
             batches.append(guids)
diff --git a/tests/test_plex.py b/tests/test_plex.py
index fcec4c4..b4fd00b 100644
--- a/tests/test_plex.py
+++ b/tests/test_plex.py
@@ -315,6 +315,24 @@ def test_guide_ids_past_the_limit_are_not_read(
     assert "more entries than are read" in caplog.text
 
 
+def test_guide_ids_are_not_read_again_after_an_error(caplog: pytest.LogCaptureFixture) -> None:
+    plex, asked = guide_by_id({n: f"g/{n}" for n in range(1, 700)})
+    get = plex._get
+
+    def failing(path: str, **params: Any) -> Any:
+        if path.endswith(",600"):
+            asked.append(path)
+            raise http.HttpError(500, path)
+        return get(path, **params)
+
+    plex._get = failing  # type: ignore[method-assign]
+    assert plex.guide_target(GUIDE, SHOW, "g/100") == "100"
+    asked.clear()
+    assert plex.guide_target(GUIDE, SHOW, "g/450") is None
+    assert len(asked) == 1
+    assert "could not all be read" in caplog.text
+
+
 def test_guide_target_is_none_for_an_entry_without_a_guid() -> None:
     plex, asked = guide_by_id({1: ""})
     assert plex.guide_target(GUIDE, SHOW, "") is None