Retry docker compose up to ride out Docker Hub pull flakes

.github/workflows/analyst-ci.yml

@@ -36,7 +36,8 @@ jobs:
         run: cp .env.example .env
 
       - name: Start the lakehouse (MinIO + REST catalog + Trino)
-        run: docker compose up -d --build minio minio-init iceberg-rest trino
+        run: |
+          for i in 1 2 3; do if docker compose up -d --build minio minio-init iceberg-rest trino; then break; fi; if [ "$i" = 3 ]; then echo "compose up still failing after $i attempts"; exit 1; fi; echo "compose up failed (attempt $i; likely a Docker Hub pull flake); retrying in 15s ..."; sleep 15; done
 
       - name: Wait for the Iceberg REST catalog
         run: |
@@ -77,7 +78,8 @@ jobs:
         run: cp .env.example .env
 
       - name: Start lakehouse + Superset (and Spark to seed)
-        run: docker compose -f docker-compose.yml -f docker-compose.services.yml up -d --build superset spark
+        run: |
+          for i in 1 2 3; do if docker compose -f docker-compose.yml -f docker-compose.services.yml up -d --build superset spark; then break; fi; if [ "$i" = 3 ]; then echo "compose up still failing after $i attempts"; exit 1; fi; echo "compose up failed (attempt $i; likely a Docker Hub pull flake); retrying in 15s ..."; sleep 15; done
 
       - name: Wait for the Iceberg REST catalog
         run: |

.github/workflows/atlas-ci.yml

@@ -39,7 +39,8 @@ jobs:
 
       # Atlas (embedded HBase + Solr) is memory-hungry — boot it alone.
       - name: Start Atlas
-        run: docker compose -f docker-compose.yml -f docker-compose.governance.yml up -d atlas
+        run: |
+          for i in 1 2 3; do if docker compose -f docker-compose.yml -f docker-compose.governance.yml up -d atlas; then break; fi; if [ "$i" = 3 ]; then echo "compose up still failing after $i attempts"; exit 1; fi; echo "compose up failed (attempt $i; likely a Docker Hub pull flake); retrying in 15s ..."; sleep 15; done
 
       - name: Wait for Atlas REST (up to ~15 min)
         run: |
@@ -61,7 +62,8 @@ jobs:
       # block the PR. It's validated locally (OpenLineage capture + the bridge).
       - name: Start Spark (+ iceberg-rest + minio)
         continue-on-error: true
-        run: docker compose -f docker-compose.yml -f docker-compose.governance.yml up -d spark
+        run: |
+          for i in 1 2 3; do if docker compose -f docker-compose.yml -f docker-compose.governance.yml up -d spark; then break; fi; if [ "$i" = 3 ]; then echo "compose up still failing after $i attempts"; exit 1; fi; echo "compose up failed (attempt $i; likely a Docker Hub pull flake); retrying in 15s ..."; sleep 15; done
 
       - name: Automatic lineage — capture from Spark (OpenLineage) into Atlas
         continue-on-error: true

.github/workflows/console-control-ci.yml

@@ -37,7 +37,8 @@ jobs:
         run: cp .env.example .env
 
       - name: Build and start lakehouse + governance + console
-        run: docker compose up -d --build console-api spark trino
+        run: |
+          for i in 1 2 3; do if docker compose up -d --build console-api spark trino; then break; fi; if [ "$i" = 3 ]; then echo "compose up still failing after $i attempts"; exit 1; fi; echo "compose up failed (attempt $i; likely a Docker Hub pull flake); retrying in 15s ..."; sleep 15; done
 
       - name: Wait for Trino
         run: |

.github/workflows/flow-ci.yml

@@ -23,7 +23,8 @@ jobs:
       - name: Create .env from example
         run: cp .env.example .env
       - name: Start NiFi
-        run: docker compose -f docker-compose.yml -f docker-compose.services.yml up -d nifi
+        run: |
+          for i in 1 2 3; do if docker compose -f docker-compose.yml -f docker-compose.services.yml up -d nifi; then break; fi; if [ "$i" = 3 ]; then echo "compose up still failing after $i attempts"; exit 1; fi; echo "compose up failed (attempt $i; likely a Docker Hub pull flake); retrying in 15s ..."; sleep 15; done
       - name: Build + run an automated flow via the REST API
         run: NIFI_URL=http://localhost:8095 bash ../platform/flow/nifi-flow.sh
       - name: Dump logs on failure

.github/workflows/governance-ci.yml

@@ -34,7 +34,8 @@ jobs:
         run: cp .env.example .env
 
       - name: Build and start lakehouse + governance
-        run: docker compose -f docker-compose.yml -f docker-compose.governance.yml up -d --build
+        run: |
+          for i in 1 2 3; do if docker compose -f docker-compose.yml -f docker-compose.governance.yml up -d --build; then break; fi; if [ "$i" = 3 ]; then echo "compose up still failing after $i attempts"; exit 1; fi; echo "compose up failed (attempt $i; likely a Docker Hub pull flake); retrying in 15s ..."; sleep 15; done
 
       - name: Wait for the Iceberg REST catalog
         run: |

.github/workflows/jupyterhub-ci.yml

@@ -29,7 +29,8 @@ jobs:
       - uses: actions/checkout@v4
 
       - name: Build and start JupyterHub
-        run: docker compose -f docker-compose.jupyterhub.yml up -d --build
+        run: |
+          for i in 1 2 3; do if docker compose -f docker-compose.jupyterhub.yml up -d --build; then break; fi; if [ "$i" = 3 ]; then echo "compose up still failing after $i attempts"; exit 1; fi; echo "compose up failed (attempt $i; likely a Docker Hub pull flake); retrying in 15s ..."; sleep 15; done
 
       - name: Smoke — hub up + spawn a notebook server
         run: bash ../platform/ml/jupyterhub/smoke.sh

.github/workflows/opdb-ci.yml

@@ -29,7 +29,8 @@ jobs:
       - uses: actions/checkout@v4
 
       - name: Start the Operational DB (HBase + Phoenix)
-        run: docker compose -f docker-compose.opdb.yml up -d
+        run: |
+          for i in 1 2 3; do if docker compose -f docker-compose.opdb.yml up -d; then break; fi; if [ "$i" = 3 ]; then echo "compose up still failing after $i attempts"; exit 1; fi; echo "compose up failed (attempt $i; likely a Docker Hub pull flake); retrying in 15s ..."; sleep 15; done
 
       - name: Phoenix smoke — create table, upsert, select
         run: bash ../platform/opdb/smoke.sh

.github/workflows/quickstart-ci.yml

@@ -34,7 +34,8 @@ jobs:
         run: docker compose config
 
       - name: Start the stack
-        run: docker compose up -d
+        run: |
+          for i in 1 2 3; do if docker compose up -d; then break; fi; if [ "$i" = 3 ]; then echo "compose up still failing after $i attempts"; exit 1; fi; echo "compose up failed (attempt $i; likely a Docker Hub pull flake); retrying in 15s ..."; sleep 15; done
 
       - name: Wait for the Iceberg REST catalog
         run: |

.github/workflows/services-ci.yml

@@ -38,7 +38,8 @@ jobs:
         run: cp .env.example .env
 
       - name: Start lakehouse + Superset (and Spark to seed)
-        run: docker compose -f docker-compose.yml -f docker-compose.services.yml up -d --build superset spark
+        run: |
+          for i in 1 2 3; do if docker compose -f docker-compose.yml -f docker-compose.services.yml up -d --build superset spark; then break; fi; if [ "$i" = 3 ]; then echo "compose up still failing after $i attempts"; exit 1; fi; echo "compose up failed (attempt $i; likely a Docker Hub pull flake); retrying in 15s ..."; sleep 15; done
 
       - name: Wait for the Iceberg REST catalog
         run: |
@@ -88,7 +89,8 @@ jobs:
         run: cp .env.example .env
 
       - name: Start the lakehouse (Trino) + Airflow
-        run: docker compose -f docker-compose.yml -f docker-compose.services.yml up -d --build trino airflow
+        run: |
+          for i in 1 2 3; do if docker compose -f docker-compose.yml -f docker-compose.services.yml up -d --build trino airflow; then break; fi; if [ "$i" = 3 ]; then echo "compose up still failing after $i attempts"; exit 1; fi; echo "compose up failed (attempt $i; likely a Docker Hub pull flake); retrying in 15s ..."; sleep 15; done
 
       - name: Wait for Trino
         run: |
@@ -144,7 +146,8 @@ jobs:
         run: cp .env.example .env
 
       - name: Start MLflow (and MinIO via its bucket init)
-        run: docker compose -f docker-compose.yml -f docker-compose.services.yml up -d --build mlflow
+        run: |
+          for i in 1 2 3; do if docker compose -f docker-compose.yml -f docker-compose.services.yml up -d --build mlflow; then break; fi; if [ "$i" = 3 ]; then echo "compose up still failing after $i attempts"; exit 1; fi; echo "compose up failed (attempt $i; likely a Docker Hub pull flake); retrying in 15s ..."; sleep 15; done
 
       - name: Wait for MLflow
         run: |
@@ -183,7 +186,8 @@ jobs:
         run: cp .env.example .env
 
       - name: Start Kafka and NiFi
-        run: docker compose -f docker-compose.yml -f docker-compose.services.yml up -d kafka nifi
+        run: |
+          for i in 1 2 3; do if docker compose -f docker-compose.yml -f docker-compose.services.yml up -d kafka nifi; then break; fi; if [ "$i" = 3 ]; then echo "compose up still failing after $i attempts"; exit 1; fi; echo "compose up failed (attempt $i; likely a Docker Hub pull flake); retrying in 15s ..."; sleep 15; done
 
       - name: Wait for the Kafka broker
         run: |

.github/workflows/spark-app-ci.yml

@@ -33,7 +33,7 @@ jobs:
         working-directory: quickstart
         run: |
           cp .env.example .env
-          docker compose up -d spark trino
+          for i in 1 2 3; do if docker compose up -d spark trino; then break; fi; if [ "$i" = 3 ]; then echo "compose up still failing after $i attempts"; exit 1; fi; echo "compose up failed (attempt $i; likely a Docker Hub pull flake); retrying in 15s ..."; sleep 15; done
 
       - name: Wait for Trino
         working-directory: quickstart

.github/workflows/streaming-ci.yml

@@ -31,7 +31,8 @@ jobs:
         run: cp .env.example .env
 
       - name: Start Kafka + Spark + the lakehouse (Trino)
-        run: docker compose -f docker-compose.yml -f docker-compose.services.yml up -d kafka spark trino
+        run: |
+          for i in 1 2 3; do if docker compose -f docker-compose.yml -f docker-compose.services.yml up -d kafka spark trino; then break; fi; if [ "$i" = 3 ]; then echo "compose up still failing after $i attempts"; exit 1; fi; echo "compose up failed (attempt $i; likely a Docker Hub pull flake); retrying in 15s ..."; sleep 15; done
 
       - name: Wait for the Kafka broker
         run: |