seed(stand): two years of activity on both stands

.github/workflows/deploy-test-stand.yml

@@ -440,14 +440,17 @@ jobs:
             echo "::warning::seeding with an explicit --email from secrets.TEST_STAND_SEED_EMAIL — the dev-lead address is NOT being read from the realm this deploy applied. Unset that secret unless this stand's realm is provisioned outside deploy/gitops."
           fi
 
-          # 365, not more: the API refuses any period ≥ 400 days.
+          # Two years, matching the compose stand. Wider than any one request
+          # the API will answer — it refuses a period of 400 days or more — so
+          # a suite asking about "the seeded period" clamps to the queryable
+          # tail (tests/stand/api/analytics/query_window).
           kube_context="$(yq -r '.kubeContext' "$INVENTORY_FILE")"
           seed_rc=0
           ./src/ingestion/tools/seed/seed-stand.sh \
             -n "$STAND_NAMESPACE" \
             --context "$kube_context" \
             "${email_args[@]}" \
-            --days 365 \
+            --days 730 \
             "${force_args[@]}" \
             > "$RUNNER_TEMP/seed.log" 2>&1 || seed_rc=$?
 

deploy/gitops/environments/test-stand/INFRA.md

@@ -141,7 +141,7 @@ not facts about any real deployment.
 |---|---|---|
 | **Fork-downgrade wedge** | Chart version resolved from `.insight-version` in a stale checkout → a downgrade; helm rewrites live Deployments into the older shape and wedges. A post-check comparing only against the *requested* version calls it a pass. | Resolver reads the OCI registry's latest tag first; the deploy workflow + `recreate-test-stand.sh` both refuse a version older than what's deployed. |
 | **`TEST_STAND_SEED_EMAIL` drift** | Dev-lead address declared in two places (realm + seeder flag) with no agreement check; only that one login fails. | `seed-stand.sh` reads the address out of the applied realm ConfigMap (keyed on the roster UUID). The flag survives only as an explicit, warned override. |
-| **730-day seed window** | Seed window wider than the analytics API's max queryable period → every window-derived request 400s, looking like a broad data problem. | Seed window capped inside the API limit; one shared query-window helper caps both suites at the same value. |
+| **730-day seed window** | Seed window wider than the analytics API's max queryable period → every window-derived request 400s, looking like a broad data problem. | Both stands seed 730 days on purpose, so the trap is live: no test may read `data_window` raw. One shared helper (`tests/stand/api/analytics/query_window`) clamps every request to the queryable tail, and a reconciliation must pass the same clamped period to both sides or it compares two different periods. |
 | **Pending-upgrade wedge** | An interrupted `helm upgrade` leaves the release `pending-upgrade`, which helm refuses to upgrade over; every later deploy fails generically. | `make deploy` checks release status first and refuses with an explicit fix; workflow step timeouts stay below helm's own `--timeout`. |
 | **Diagnostics-allowlist retreat** | The workflow published curated cluster diagnostics into the public log; an allowlist like that only ever grows. | A failed run prints only the dead stage + edge probe status codes. Full output is read operator-side. |
 | **`max_user_connections` crash-loop** | The shared MariaDB user defaulted to the operator's low per-user limit (§1). | `maxUserConnections: 100` on `User/insight`. |

docker-compose.yml

@@ -724,7 +724,11 @@ services:
       # ISO date pins the dataset reproducibly, the literal `today` (or unset)
       # tracks the calendar. Both are recorded in the emitted manifest.
       SEED_ANCHOR_DATE: "${SEED_ANCHOR_DATE:-}"
-      SEED_DAYS: "${SEED_DAYS:-60}"
+      # Two years, matching the deployed stand. Wider than any one request the
+      # analytics API will answer (it refuses a period of 400 days or more), so
+      # a test asking about "the seeded period" must clamp — see
+      # tests/stand/api/analytics/query_window.
+      SEED_DAYS: "${SEED_DAYS:-730}"
       # The issuer the authenticator resolved for this run. `test-stand up`
       # persists it after cmd_up so the manifest
       # records the real IdP instead of assuming the default.

src/ingestion/tools/seed/insight_seed/generators/git.py

@@ -121,11 +121,10 @@ def seed_class_git_commits(
     ]
     rows: list[tuple[object, ...]] = []
     version = 1
-    # Dealt to the dev lead's first commits in generation order (a merge
-    # commit is skipped — gold's evidence model filters those out, so a
-    # message on one would never reach the drilldown). Deterministic across
-    # re-seeds because the commit stream itself is.
-    hostile_messages = list(HOSTILE_COMMIT_MESSAGES)
+    # Row indices of the dev lead's non-merge commits, ascending by date. Gold's
+    # evidence model filters merge commits out, so a message on one would never
+    # reach the drilldown.
+    lead_commits: list[int] = []
     for p in _eligible(roster):
         persona = persona_multiplier(p.uuid)
         weight = TEAM_PROFILES[p.team or ""].weights["github"]
@@ -139,9 +138,8 @@ def seed_class_git_commits(
                 # LOC per commit capped at ≤200 by construction.
                 added = float(rng.randint(2, 180))
                 removed = float(rng.randint(0, 80))
-                message = ""
-                if p.uuid == DEV_LEAD_UUID and not is_merge and hostile_messages:
-                    message = hostile_messages.pop(0)
+                if p.uuid == DEV_LEAD_UUID and not is_merge:
+                    lead_commits.append(len(rows))
                 rows.append(
                     (
                         tenant_uuid,
@@ -157,10 +155,24 @@ def seed_class_git_commits(
                         added,
                         removed,
                         "insight_github",
-                        message,
+                        "",
                         version,
                     )
                 )
+
+    # The most recent of those commits, not the earliest: a suite asking about
+    # "the seeded period" asks about the tail the API will answer for, and a
+    # window wider than that cap would leave titles dealt to the oldest days
+    # unreachable.
+    message_at = cols.index("message")
+    for offset, hostile in enumerate(reversed(HOSTILE_COMMIT_MESSAGES), start=1):
+        if offset > len(lead_commits):
+            break
+        index = lead_commits[-offset]
+        row = list(rows[index])
+        row[message_at] = hostile
+        rows[index] = tuple(row)
+
     return bulk_insert(client, "silver", "class_git_commits", cols, rows)
 
 

tests/stand/api/analytics/test_drilldown.py

@@ -66,6 +66,7 @@
     MetricDrilldownColumnType,
     MetricDrilldownResponse,
 )
+from . import query_window
 from .drilldown_matrix import EXPORT_SHAPES, MATRIX, Expectation, Tier
 
 DRILLDOWN = analytics_path("/v1/metric-drilldown")
@@ -154,7 +155,7 @@ def _request_for(
     filters: Sequence[JsonValue] = (),
     display_dimensions: Sequence[str] = (),
 ) -> dict[str, JsonValue]:
-    start, _, end = manifest.data_window.partition("..")
+    start, end = query_window(manifest)
     request: dict[str, JsonValue] = {
         "metric_key": metric_key,
         "entity": {
@@ -286,7 +287,7 @@ def _period_value(
     person with no observations in the period has no value at all, and that is
     what an empty evidence page has to agree with.
     """
-    start, _, end = manifest.data_window.partition("..")
+    start, end = query_window(manifest)
     person_id = manifest.fixture("dev_lead").uuid
     response = api.post(
         METRIC_RESULTS,

tests/stand/api/analytics/test_drilldown_rebuild.py

@@ -48,6 +48,7 @@
 
 from ..schemas import MetricResultsResponse, PeriodView, ProblemDocument
 from ..schemas.analytics import MetricDrilldownResponse
+from . import query_window
 
 pytestmark = pytest.mark.reliability
 
@@ -195,7 +196,7 @@ def _remove_seed_containers(project: str) -> None:
 def _request(
     manifest: Manifest, *, limit: int | None = None, cursor: str | None = None
 ) -> dict[str, JsonValue]:
-    start, _, end = manifest.data_window.partition("..")
+    start, end = query_window(manifest)
     request: dict[str, JsonValue] = {
         "metric_key": GIT_COMMITS,
         "entity": {"type": "person", "id": manifest.fixture("dev_lead").uuid},
@@ -229,7 +230,7 @@ def _walk(api: ApiClient, manifest: Manifest) -> list[Mapping[str, Any]]:
 
 def _period_value(api: ApiClient, manifest: Manifest) -> float | None:
     """The scalar the dashboard shows for the same selection — the walk's oracle."""
-    start, _, end = manifest.data_window.partition("..")
+    start, end = query_window(manifest)
     person_id = manifest.fixture("dev_lead").uuid
     response = api.post(
         METRIC_RESULTS,