fix(pipeline): cache invalidation for IDACI DAG too; curl timeouts

Addresses AI-review findings: the annual IDACI DAG also rebuilds a mart (fact_deprivation) and needs the reload; curl gets connect/max timeouts so an unreachable backend fails fast instead of hanging the task. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
fix(pipeline): actually invalidate the backend cache after data rebuilds
2026-07-09 21:26:45 +01:00 · 2026-07-09 21:12:01 +01:00
1 changed files with 49 additions and 4 deletions
@@ -38,6 +38,31 @@ default_args = {
    "retry_delay": timedelta(minutes=5),
 }

+# The backend caches the marts DataFrame at startup; after any rebuild the
+# cache must be invalidated or the API serves stale (or empty) data until the
+# container restarts.
+INVALIDATE_CACHE_CMD = """
+set -e
+BACKEND_URL="${BACKEND_URL:-http://backend:80}"
+ADMIN_KEY="${ADMIN_API_KEY:-changeme}"
+
+echo "Calling $BACKEND_URL/api/admin/reload ..."
+
+response=$(curl -s -o /tmp/reload_response.json -w "%{http_code}" \\
+    --connect-timeout 10 --max-time 120 \\
+    -X POST "$BACKEND_URL/api/admin/reload" \\
+    -H "X-API-Key: $ADMIN_KEY" \\
+    -H "Content-Type: application/json")
+
+echo "HTTP status: $response"
+cat /tmp/reload_response.json
+
+if [ "$response" != "200" ]; then
+    echo "ERROR: backend cache reload failed (HTTP $response)"
+    exit 1
+fi
+"""
+

 # ── Daily DAG (GIAS + downstream) ──────────────────────────────────────

@@ -91,7 +116,12 @@ print(f'Validation passed: {{count}} GIAS rows')
        bash_command=f"cd {PIPELINE_DIR} && python scripts/sync_typesense.py",
    )

-    extract_group >> validate_raw >> dbt_build >> sync_typesense
+    invalidate_cache = BashOperator(
+        task_id="invalidate_cache",
+        bash_command=INVALIDATE_CACHE_CMD,
+    )
+
+    extract_group >> validate_raw >> dbt_build >> sync_typesense >> invalidate_cache


 # ── Monthly DAG (Ofsted) ───────────────────────────────────────────────
@@ -121,7 +151,12 @@ with DAG(
        bash_command=f"cd {PIPELINE_DIR} && python scripts/sync_typesense.py",
    )

-    extract_ofsted >> dbt_build_ofsted >> sync_typesense_ofsted
+    invalidate_cache_ofsted = BashOperator(
+        task_id="invalidate_cache",
+        bash_command=INVALIDATE_CACHE_CMD,
+    )
+
+    extract_ofsted >> dbt_build_ofsted >> sync_typesense_ofsted >> invalidate_cache_ofsted


 # ── Annual DAG (EES: KS2, KS4, Census, Admissions) ───────────────────
@@ -153,7 +188,12 @@ with DAG(
        bash_command=f"cd {PIPELINE_DIR} && python scripts/sync_typesense.py",
    )

-    extract_ees_group >> dbt_build_ees >> sync_typesense_ees
+    invalidate_cache_ees = BashOperator(
+        task_id="invalidate_cache",
+        bash_command=INVALIDATE_CACHE_CMD,
+    )
+
+    extract_ees_group >> dbt_build_ees >> sync_typesense_ees >> invalidate_cache_ees


 # ── Annual DAG (IDACI Deprivation) ────────────────────────────────────
@@ -178,4 +218,9 @@ with DAG(
        bash_command=f"cd {PIPELINE_DIR}/transform && {DBT_BIN} build --profiles-dir . --target production --select stg_idaci+ fact_deprivation+",
    )

-    extract_idaci >> dbt_build_idaci
+    invalidate_cache_idaci = BashOperator(
+        task_id="invalidate_cache",
+        bash_command=INVALIDATE_CACHE_CMD,
+    )
+
+    extract_idaci >> dbt_build_idaci >> invalidate_cache_idaci
Author	SHA1	Message	Date
TudorandClaude Fable 5	d677b54533	fix(pipeline): cache invalidation for IDACI DAG too; curl timeouts PR Checks / Frontend Typecheck + Tests (pull_request) Successful in 9m37s Details PR Checks / Backend Smoke (pull_request) Successful in 6s Details PR Checks / Build Backend (no push) (pull_request) Successful in 18s Details PR Checks / Build Frontend (no push) (pull_request) Successful in 48s Details PR Checks / Build Pipeline (no push) (pull_request) Successful in 35s Details PR Checks / AI Code Review (Claude) (pull_request) Successful in 1m40s Details Addresses AI-review findings: the annual IDACI DAG also rebuilds a mart (fact_deprivation) and needs the reload; curl gets connect/max timeouts so an unreachable backend fails fast instead of hanging the task. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>	2026-07-09 21:26:45 +01:00
TudorandClaude Fable 5	c353e36072	fix(pipeline): actually invalidate the backend cache after data rebuilds PR Checks / Frontend Typecheck + Tests (pull_request) Successful in 9m40s Details PR Checks / Backend Smoke (pull_request) Successful in 6s Details PR Checks / Build Backend (no push) (pull_request) Successful in 16s Details PR Checks / Build Frontend (no push) (pull_request) Successful in 51s Details PR Checks / Build Pipeline (no push) (pull_request) Successful in 36s Details PR Checks / AI Code Review (Claude) (pull_request) Failing after 1m20s Details The daily/monthly/annual DAG docstring promised an Invalidate Cache step that never existed — after a marts rebuild the backend kept serving its startup-cached (possibly empty) DataFrame until a container restart. Add a POST /api/admin/reload task at the end of each pipeline DAG, mirroring the sitemap DAG's admin-call pattern. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>	2026-07-09 21:12:01 +01:00