Harden CI deploy timeouts for heavier post-upgrade images

.github/workflows/ci.yml

@@ -339,7 +339,7 @@ jobs:
           cd terraform
           terraform init -backend=false
           terraform apply -var-file=environments/local.tfvars -auto-approve
-        timeout-minutes: 20
+        timeout-minutes: 30
 
       - name: Wait for Airflow database migrations
         run: |

.github/workflows/platform-tests.yml

@@ -1,7 +1,7 @@
 ---
 name: Platform Tests
 
 on:
   push:
     branches: [main]
   pull_request:
@@ -369,7 +369,7 @@
         run: |
           cd terraform
           terraform apply -var-file=environments/local.tfvars -auto-approve
-        timeout-minutes: 15
+        timeout-minutes: 30
 
       - name: Wait for Pods
         run: |

.gitignore

@@ -161,4 +161,6 @@ build/
 
 # Local development
 local/
-scratch/
+scratch/
+# Claude Code harness state
+.claude/

terraform/main.tf

@@ -74,13 +74,23 @@ module "spark" {
 resource "kubernetes_deployment_v1" "jupyter" {
   depends_on = [module.spark, module.minio]
 
+  # The jupyter/pyspark-notebook image is large (~4GB); on a constrained
+  # node (e.g. minikube in CI) the initial pull can exceed the provider's
+  # default 10m rollout wait. Give it more headroom. See issue #36 for the
+  # durable fix (pinning a smaller image).
+  timeouts {
+    create = "20m"
+    update = "20m"
+  }
+
   metadata {
     name      = "jupyter"
     namespace = kubernetes_namespace_v1.ldp.metadata[0].name
   }
 
   spec {
-    replicas = 1
+    replicas                  = 1
+    progress_deadline_seconds = 1200
 
     selector {
       match_labels = {

terraform/modules/airflow/main.tf

@@ -76,7 +76,7 @@ resource "helm_release" "airflow" {
     EOT
   ]
 
-  timeout       = 900 # Increased for CI environment (15 minutes)
+  timeout       = 1500 # Increased for CI image-pull contention (25 minutes)
   wait          = true
   wait_for_jobs = true
 }

terraform/modules/monitoring/main.tf

@@ -1,5 +1,12 @@
 # Prometheus deployment for metrics collection
 resource "kubernetes_deployment_v1" "prometheus" {
+  # On a constrained node (e.g. minikube in CI) image-pull contention can
+  # push the rollout past the provider's default 10m wait. Give it headroom.
+  timeouts {
+    create = "20m"
+    update = "20m"
+  }
+
   metadata {
     name      = "prometheus"
     namespace = var.namespace
@@ -9,7 +16,8 @@ resource "kubernetes_deployment_v1" "prometheus" {
   }
 
   spec {
-    replicas = 1
+    replicas                  = 1
+    progress_deadline_seconds = 1200
 
     selector {
       match_labels = {
@@ -194,6 +202,13 @@ resource "kubernetes_config_map_v1" "prometheus_config" {
 
 # Grafana deployment for visualization
 resource "kubernetes_deployment_v1" "grafana" {
+  # On a constrained node (e.g. minikube in CI) image-pull contention can
+  # push the rollout past the provider's default 10m wait. Give it headroom.
+  timeouts {
+    create = "20m"
+    update = "20m"
+  }
+
   metadata {
     name      = "grafana"
     namespace = var.namespace
@@ -203,7 +218,8 @@ resource "kubernetes_deployment_v1" "grafana" {
   }
 
   spec {
-    replicas = 1
+    replicas                  = 1
+    progress_deadline_seconds = 1200
 
     selector {
       match_labels = {