diff --git a/prisma/migrations/0077_v1447_drop_legacy_consecutive_failures/migration.sql b/prisma/migrations/0077_v1447_drop_legacy_consecutive_failures/migration.sql
new file mode 100644
index 000000000..d7f134683
--- /dev/null
+++ b/prisma/migrations/0077_v1447_drop_legacy_consecutive_failures/migration.sql
@@ -0,0 +1,49 @@
+-- v1.4.47 W1 — drop the legacy `consecutive_failures` column on
+-- `integration_statuses`.
+--
+-- v1.4.43 W14 introduced `consecutive_failures_by_kind` (JSONB) as
+-- the bucketed per-FailureKind counter and kept the single-column
+-- `consecutive_failures` integer for one release as a fallback so
+-- v1.4.43 readers wouldn't break mid-deploy. v1.4.45 W10 closed the
+-- write-time gap (back-fill increment posts the bucket and the legacy
+-- column at the same N+1 value). v1.4.47 W1 drops the legacy column
+-- now that every active integration in production has written at
+-- least one new failure or success since v1.4.43 shipped (>24h ago),
+-- so every row's `consecutive_failures_by_kind` is populated and the
+-- legacy integer is redundant.
+--
+-- After this migration the alert-ladder reader uses
+-- `Math.max(...Object.values(buckets))` exclusively — the legacy
+-- column is no longer consulted. `recordSyncFailure` and friends
+-- stop writing to `consecutive_failures` entirely (the corresponding
+-- application-level cleanup in `src/lib/integrations/status.ts`
+-- lands in the same release).
+--
+-- Idempotent guard (`IF EXISTS`) matches the 0075 pattern so reruns
+-- against an environment that has already dropped the column are
+-- safe.
+--
+-- Reversibility (down-migration):
+--
+--   ALTER TABLE "integration_statuses"
+--       ADD COLUMN IF NOT EXISTS "consecutive_failures" INTEGER
+--       NOT NULL DEFAULT 0;
+--
+--   UPDATE "integration_statuses"
+--      SET "consecutive_failures" = COALESCE(
+--          GREATEST(
+--              ("consecutive_failures_by_kind"->>'transient')::int,
+--              ("consecutive_failures_by_kind"->>'reauth_required')::int,
+--              ("consecutive_failures_by_kind"->>'persistent')::int
+--          ),
+--          0
+--      )
+--      WHERE "consecutive_failures_by_kind" IS NOT NULL;
+--
+-- The down-script restores the legacy counter to `Math.max(...buckets)`
+-- — the same value the alert ladder reads from the buckets — so a
+-- rollback lands every row at the post-v1.4.43 counter value without
+-- losing the streak's true age.
+
+ALTER TABLE "integration_statuses"
+    DROP COLUMN IF EXISTS "consecutive_failures";
diff --git a/prisma/schema.prisma b/prisma/schema.prisma
index 5c94f1d9b..e1a03e68e 100644
--- a/prisma/schema.prisma
+++ b/prisma/schema.prisma
@@ -1374,11 +1374,12 @@ model MedicationComplianceRollup {
 //   - consecutive failure counter (resets on success)
 //
 // `recordSyncSuccess` / `recordSyncFailure` helpers in
-// `src/lib/integrations/status.ts` are the only writers. The
-// `consecutiveFailures` counter feeds the admin-Telegram alerting
-// threshold (default 3). State `error_reauth` parks scheduled syncs
-// until the user reconnects (the sync entry-points consult this row
-// before doing any external work).
+// `src/lib/integrations/status.ts` are the only writers. The per-kind
+// `consecutiveFailuresByKind` JSON bucket feeds the admin-Telegram
+// alerting threshold (default 3) via `Math.max(...Object.values(...))`.
+// State `error_reauth` parks scheduled syncs until the user reconnects
+// (the sync entry-points consult this row before doing any external
+// work).
 //
 // `integration` values are bare strings ("withings", "moodlog") rather
 // than an enum so future integrations don't require a migration.
@@ -1390,7 +1391,9 @@ model IntegrationStatus {
   lastSuccessAt              DateTime? @map("last_success_at")
   lastAttemptAt              DateTime? @map("last_attempt_at")
   lastError                  String?   @map("last_error") // AES-256-GCM encrypted (may contain token-fragment metadata)
-  consecutiveFailures        Int       @default(0) @map("consecutive_failures") // v1.4.44 will drop; per-kind counter below is the source of truth
+  // v1.4.47 W1 — legacy single-column `consecutiveFailures` integer was
+  // dropped (migration 0076). The per-kind bucket below is now the
+  // sole source of truth for the alert ladder.
   consecutiveFailuresByKind  Json?     @map("consecutive_failures_by_kind") // v1.4.43 W14 — { transient, reauth_required, persistent }
   persistentFailureStartedAt DateTime? @map("persistent_failure_started_at") // v1.4.43 W14 — start of the current persistent streak (drives >24h park)
   alertedAt                  DateTime? @map("alerted_at") // last time admin Telegram alert fired (idempotency guard)