fix: Make Chronon demo setup reproducible and surface startup failures

Signed-off-by: Francisco Javier Arceo <arceofrancisco@gmail.com>
franciscojavierarceo · Oct 5, 2026 · Jun 22, 2026 · Jun 22, 2026 · Jun 22, 2026 · Jun 22, 2026
commit eae98902b0d76bad50592a2d592a3760a1bb65bd
diff --git a/.github/workflows/pr_chronon_integration_tests.yml b/.github/workflows/pr_chronon_integration_tests.yml
@@ -102,7 +102,7 @@ jobs:
           sudo mkdir -p /usr/src/thrift
           sudo tar zxf /tmp/thrift.tar.gz -C /usr/src/thrift --strip-components=1
           cd /usr/src/thrift
-          sudo ./configure --without-python --without-cpp
+          sudo ./configure --disable-libs --disable-tests --disable-tutorial
           sudo make -j2
           sudo make install
           python3 -m pip install --break-system-packages build

diff --git a/README.md b/README.md
@@ -173,6 +173,7 @@ The list below contains the functionality that contributors are planning to deve
   * [x] Vector Search (Alpha release. See [RFC](https://docs.google.com/document/d/18IWzLEA9i2lDWnbfbwXnMCg3StlqaLVI-uRpQjr_Vos/edit#heading=h.9gaqqtox9jg6))
   * [ ] [Enhanced Feature Server and SDK for native support for NLP](https://github.com/feast-dev/feast/issues/4964)
 * **Data Sources**
+  * [x] [Chronon source (contrib)](https://docs.feast.dev/reference/data-sources/chronon)
   * [x] [Snowflake source](https://docs.feast.dev/reference/data-sources/snowflake)
   * [x] [Redshift source](https://docs.feast.dev/reference/data-sources/redshift)
   * [x] [BigQuery source](https://docs.feast.dev/reference/data-sources/bigquery)
@@ -189,6 +190,7 @@ The list below contains the functionality that contributors are planning to deve
   * [x] [Ray source (contrib plugin)](https://docs.feast.dev/reference/data-sources/ray)
   * [x] Kafka / Kinesis sources (via [push support into the online store](https://docs.feast.dev/reference/data-sources/push))
 * **Offline Stores**
+  * [x] [Chronon (contrib)](https://docs.feast.dev/reference/offline-stores/chronon)
   * [x] [Snowflake](https://docs.feast.dev/reference/offline-stores/snowflake)
   * [x] [Redshift](https://docs.feast.dev/reference/offline-stores/redshift)
   * [x] [BigQuery](https://docs.feast.dev/reference/offline-stores/bigquery)
@@ -209,6 +211,7 @@ The list below contains the functionality that contributors are planning to deve
   * [x] [Hybrid](https://docs.feast.dev/reference/offline-stores/hybrid)
   * [x] [Custom offline store support](https://docs.feast.dev/how-to-guides/customizing-feast/adding-a-new-offline-store)
 * **Online Stores**
+  * [x] [Chronon](https://docs.feast.dev/reference/online-stores/chronon)
   * [x] [Snowflake](https://docs.feast.dev/reference/online-stores/snowflake)
   * [x] [DynamoDB](https://docs.feast.dev/reference/online-stores/dynamodb)
   * [x] [Redis](https://docs.feast.dev/reference/online-stores/redis)

@@ -6,6 +6,10 @@ Please see [Offline Store](../../getting-started/components/offline-store.md) fo
 [overview.md](overview.md)
 {% endcontent-ref %}
 
+{% content-ref url="chronon.md" %}
+[chronon.md](chronon.md)
+{% endcontent-ref %}
+
 {% content-ref url="dask.md" %}
 [dask.md](dask.md)
 {% endcontent-ref %}

@@ -29,6 +29,11 @@ There are currently four core offline store implementations: `DaskOfflineStore`,
 There are several additional implementations contributed by the Feast community  (`PostgreSQLOfflineStore`, `SparkOfflineStore`, `TrinoOfflineStore`, and `RayOfflineStore`), which are not guaranteed to be stable or to match the functionality of the core implementations.
 Details for each specific offline store, such as how to configure it in a `feature_store.yaml`, can be found [here](README.md).
 
+The community [Chronon offline store](chronon.md) reads materialized Parquet with
+point-in-time joins and supports local on-demand transforms and Parquet saved
+datasets. Computation and materialization remain managed by Chronon; see its
+[functionality matrix](chronon.md#functionality-matrix) for supported operations.
+
 Below is a matrix indicating which offline stores support which methods.
 
 || | Dask | BigQuery | Snowflake | Redshift | Postgres | Spark | Trino | Couchbase | Ray |
@@ -55,4 +60,4 @@ Below is a matrix indicating which `RetrievalJob`s support what functionality.
 || remote execution of Python-based on-demand transforms | no  | no  | no  | no  | no  | no  | no  | no  | no  | no  |
 || persist results in the offline store                  | yes | yes | yes | yes | yes | yes | no  | yes | yes | yes |
 || preview the query plan before execution               | yes | yes | yes | yes | yes | yes | yes | no  | yes | yes |
-|| read partitioned data                                 | yes | yes | yes | yes | yes | yes | yes | yes | yes | yes |
+|| read partitioned data                                 | yes | yes | yes | yes | yes | yes | yes | yes | yes | yes |
@@ -6,6 +6,10 @@ Please see [Online Store](../../getting-started/components/online-store.md) for
 [overview.md](overview.md)
 {% endcontent-ref %}
 
+{% content-ref url="chronon.md" %}
+[chronon.md](chronon.md)
+{% endcontent-ref %}
+
 {% content-ref url="sqlite.md" %}
 [sqlite.md](sqlite.md)
 {% endcontent-ref %}

@@ -13,6 +13,11 @@ uv run python examples/chronon/run_demo.py --offline-only
 
 ## Live Chronon Service
 
+The pinned quickstart Docker image is ARM64. Use an ARM64 Docker host (as in
+the integration workflow), or provide a compatible image for your host with
+`CHRONON_QUICKSTART_IMAGE`. Docker must be running and accessible to your user.
+The launcher uses MongoDB 7.0; `CHRONON_MONGO_IMAGE` overrides that image.
+
 Install `sbt` and a JDK. On macOS with Homebrew, this is enough:
 
 ```bash

@@ -14,6 +14,8 @@ SERVICE_PID_FILE="${CHRONON_SERVICE_PID_FILE:-/tmp/chronon-service.pid}"
 SERVICE_LOG_FILE="${CHRONON_SERVICE_LOG_FILE:-/tmp/chronon-service.log}"
 JAVA_BIN="${JAVA_BIN:-java}"
 PYTHON_BIN="${PYTHON_BIN:-python3}"
+QUICKSTART_IMAGE="${CHRONON_QUICKSTART_IMAGE:-ezvz/chronon@sha256:a3f6b5e1aa85bb23269ebd67fc7e4b0fb2d0cf3e53bc2a45abdc23c95769415c}"
+MONGO_IMAGE="${CHRONON_MONGO_IMAGE:-mongo:7.0@sha256:1f995ad6fdb93244a1addab1b58f934a0bc2f5643c38e02f5e9d7f0c7d227a7b}"
 CHRONON_PREFLIGHT_ONLY="${CHRONON_PREFLIGHT_ONLY:-0}"
 MONGO_IMPL_JAR="${CHRONON_DIR}/quickstart/mongo-online-impl/target/scala-2.12/mongo-online-impl-assembly-0.1.0-SNAPSHOT.jar"
 SERVICE_TARGET_DIR="${CHRONON_DIR}/service/target/scala-2.12"
@@ -94,6 +96,11 @@ cleanup_stale_service() {
 wait_for_mongo() {
   local attempts=60
   until docker exec "${MONGO_CONTAINER}" mongosh --quiet --eval 'db.runCommand({ ping: 1 }).ok' >/dev/null 2>&1; do
+    if [[ "$(docker inspect --format '{{.State.Running}}' "${MONGO_CONTAINER}")" != "true" ]]; then
+      echo "Mongo container exited before becoming ready." >&2
+      docker logs "${MONGO_CONTAINER}" >&2 || true
+      exit 1
+    fi
     attempts=$((attempts - 1))
     if [[ "${attempts}" -le 0 ]]; then
       echo "Mongo did not become ready in time." >&2
@@ -163,7 +170,7 @@ docker run -d \
   -p 27017:27017 \
   -e MONGO_INITDB_ROOT_USERNAME=admin \
   -e MONGO_INITDB_ROOT_PASSWORD=admin \
-  mongo:latest >/dev/null
+  "${MONGO_IMAGE}" >/dev/null
 
 wait_for_mongo
 
@@ -184,7 +191,7 @@ docker run -d \
   -e CHRONON_ONLINE_CLASS=ai.chronon.quickstart.online.ChrononMongoOnlineImpl \
   -e "CHRONON_ONLINE_ARGS=-Zuser=admin -Zpassword=admin -Zhost=${MONGO_CONTAINER} -Zport=27017 -Zdatabase=admin" \
   -v "${CHRONON_DIR}/quickstart/mongo-online-impl:/srv/onlineImpl" \
-  ezvz/chronon \
+  "${QUICKSTART_IMAGE}" \
   bash -lc '/opt/spark/bin/spark-shell -i scripts/data-loader.scala && touch /tmp/chronon-data-load.complete && tail -f /dev/null' >/dev/null
 
 wait_for_data_load

@@ -43,9 +43,9 @@ def test_start_script_can_fail_before_docker_when_repo_missing(tmp_path: Path):
     assert "git clone" in result.stderr
 
 
-@pytest.mark.parametrize("loader_fails", [False, True])
-def test_start_script_waits_for_completed_data_load(
-    tmp_path: Path, loader_fails: bool
+@pytest.mark.parametrize("failed_container", ["", "chronon-main", "chronon-mongo"])
+def test_start_script_requires_ready_containers(
+    tmp_path: Path, failed_container: str
 ) -> None:
     chronon_repo = tmp_path / "chronon"
     jar = (
@@ -68,17 +68,20 @@ def test_start_script_waits_for_completed_data_load(
 
 args = sys.argv[1:]
 state = Path(os.environ['TEST_DOCKER_STATE'])
-failed = os.environ['TEST_LOADER_FAILS'] == '1'
+failed_container = os.environ['TEST_FAILED_CONTAINER']
+failed = failed_container == 'chronon-main'
 if args[0] == 'run' and 'chronon-main' in args:
     state.write_text('0')
+elif args[:2] == ['exec', 'chronon-mongo']:
+    sys.exit(1 if failed_container == 'chronon-mongo' else 0)
 elif args[:2] == ['logs', 'chronon-main']:
     print("Spark session available as 'spark'.")
 elif args[:3] == ['exec', 'chronon-main', 'test']:
     polls = int(state.read_text()) + 1
     state.write_text(str(polls))
     sys.exit(0 if polls >= 2 and not failed else 1)
 elif args[0] == 'inspect':
-    print('false' if failed else 'true')
+    print('false' if args[-1] == failed_container else 'true')
 elif args[:3] == ['exec', 'chronon-main', 'bash']:
     if failed or int(state.read_text()) < 2:
         print('Upload started before successful data load', file=sys.stderr)
@@ -103,17 +106,22 @@ def test_start_script_waits_for_completed_data_load(
             "CHRONON_PREFLIGHT_ONLY": "0",
             "PYTHON_BIN": "/usr/bin/true",
             "TEST_DOCKER_STATE": str(state),
-            "TEST_LOADER_FAILS": "1" if loader_fails else "0",
+            "TEST_FAILED_CONTAINER": failed_container,
         },
         capture_output=True,
         text=True,
         timeout=15,
         check=False,
     )
-    if loader_fails:
+    if failed_container:
         assert result.returncode == 1
         assert not state.with_suffix(".uploaded").exists()
-        assert "data loader exited" in result.stderr
+        expected = (
+            "data loader exited"
+            if failed_container == "chronon-main"
+            else "Mongo container exited"
+        )
+        assert expected in result.stderr
     else:
         assert result.returncode == 0, result.stderr
         assert state.with_suffix(".uploaded").exists()