@leverege/build-tools 2.68.3-beta.4 → 2.69.1

package/.vscode/build-tools.code-workspace

@@ -0,0 +1,8 @@
+{
+	"folders": [
+		{
+			"path": "."
+		}
+	],
+	"settings": {}
+}

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@leverege/build-tools",
-  "version": "2.68.3-beta.4",
+  "version": "2.69.1",
   "description": "A collection of build / support tools for Leverege developers",
   "main": "index.js",
   "repository": {
@@ -86,7 +86,7 @@
     "shell-quote": "^1.8.3",
     "simple-git": "^3.28.0",
     "toml": "^3.0.0",
-    "zx": "^8.6.1"
+    "zx": "^8.7.1"
   },
   "devDependencies": {
     "@leverege/eslint-config-leverege": "^5.0.1",

package/src/docker-to-registry.mjs

@@ -14,11 +14,13 @@ import {
   getGitBranchAndUpstream,
   gitRepoIsDirty,
   proceed,
-  shellCmd } from './Utils.mjs'
+  shellCmd,
+  warning,
+} from './Utils.mjs'
 
 import docker from './Docker.mjs'
 
-const minimumBuildToolsVersion = '2.68.0'
+const minimumBuildToolsVersion = '2.69.0'
 
 // refresh-npm-token does not throw so no need to try, but it emits in debug
 const refreshErr = await shellCmd( 'refresh-npm-token' )
@@ -91,7 +93,8 @@ const repoDescr = await analyzeRepository( giveGuidance )
 debug( { repoDescr }, '<==The Repo Description' )
 
 if ( semver.lt( semver.coerce( repoDescr.buildToolsVersion ), minimumBuildToolsVersion ) ) {
-  errorExit( `Error: package.json @leverege/build-tools minimum version is ^${minimumBuildToolsVersion}` )
+  warning( `package.json @leverege/build-tools preferred minimum version is ^${minimumBuildToolsVersion}\n` )
+  await proceed()
 }
 
 const {

package/src/helm-audit.mjs

@@ -12,6 +12,12 @@ const spinner = ora( 'Gathering Helm release info...' ).start()
 const limit = pLimit( 5 ) // limit concurrent shellCmds to avoid API overload
 const OMIT_NAMESPACES = new Set( [ 'estafette' ] )
 
+process.on( 'SIGINT', () => {
+  // catch ctrl^C in order to reset terminal back to normal
+  spinner.stop()  
+  process.exit( 1 )
+} )
+
 async function getHelmReleases() {
   const json = await shellCmd( 'helm list -A -o json' )
   return JSON.parse( json ).filter( r => !OMIT_NAMESPACES.has( r.namespace ) )

package/src/helm-charts/geotile-server/values-local.yaml

@@ -6,5 +6,10 @@ config:
   UPLOADER_PROJECT_ID: "${PROJECT_ID}"
   UPLOADER_BUCKET: "${PROJECT_ID}-geotile-server"
 
+  # CNPG override
+  SQL_HOST: "cnpg-db-psql-stack-rw.cnpg-operands" # vs postgres-postgresql
+  SQL_PASSWORD_SECRET: "cnpg-db-psql-stack-postgres-pw"
+  SQL_PASSWORD_KEY: "password"
+
 serviceMonitor:
   enabled: true

package/src/helm-charts/prom-operator/README-GPU.md

@@ -0,0 +1,42 @@
+# 📌 Optional Prometheus Custom Metrics
+
+## 🎯 Why we do this
+We maintain a **single, consistent Prometheus baseline** for all clusters. Certain metrics — like GPU duty cycle percent or other node-type-specific stats — **do not exist in every cluster**. If we deploy scrape configs or custom adapter rules for optional metrics globally, we risk:
+
+- Empty series queries → scrape flapping.
+- Duplicate or rejected samples.
+- Confusing "phantom target down" or noisy alerts.
+
+## ✅ How we solve it
+- **Baseline:**  
+  Our base `prom-operator` Helm chart and Prometheus Adapter configs **do not contain cluster-specific custom metrics** by default.
+  - running _helmup prom-operator_ will deploy prometheus with the standard exporters
+
+- **Specialized nodes:**  
+  When clusters require optional metrics (e.g., GPU nodes, special sensors, vendor-specific exporters), we deploy those configs with a separate, explicit step:
+  - post helmup run _./prom-operator/gpu-hpa-metrics.sh_ to apply the prometheus-adapter.yaml which contains the GPU scraper
+
+## 🔐 Benefits
+- Zero config drift: our core charts stay generic and versioned.
+- No surprise scrape issues or prometheus TSDB rejects in clusters that don’t expose those metrics.
+- Clusters only carry the scrape configs they can actually fulfill.
+
+## 🛠️ Operational tips
+- **Add custom metrics only when a cluster’s node pool truly provides them.**
+- Keep custom metric configs organized in a clear, dedicated folder (`gpu-adapter-values.yaml`, `gpu-metrics.yaml`, etc.)
+- When adding new clusters, verify whether specialized nodes are present and apply the custom metrics if needed.
+- When GPU adoption grows, we can promote the configs to the baseline and gate them behind a `gpuEnabled` flag.
+
+## ⚙️ Example
+> Our `gpu_duty_cycle_percent` metric is installed *only* on clusters with GPU node pools.  
+> This is done by running:
+>
+> ```bash
+> ./prom-operator/gpu-hpa-metrics.sh
+> ```
+> or directly with helm:
+> ```bash
+> helm upgrade prometheus-adapter -f gpu-adapter-values.yaml
+> ```
+>
+> after the base Prometheus stack is installed.

package/src/helm-charts/prom-operator/gpu-hpa-metrics.sh

@@ -0,0 +1,12 @@
+#!/bin/bash
+
+# https://artifacthub.io/packages/helm/prometheus-community/prometheus-adapter
+showInstalling "The Prometheus Adapter (prom-operator custom HPA metrics)"
+[ -z "$PROMETHEUS_ADAPTER_CHART_VERSION" ] && PROMETHEUS_ADAPTER_CHART_VERSION="4"
+helm upgrade $NS --install prometheus-adapter prometheus-community/prometheus-adapter \
+                 --values prom-operator/prometheus-adapter.yaml \
+                 --version $PROMETHEUS_ADAPTER_CHART_VERSION $HELM_WHAT
+
+# finally, apply the PrometheusRule to generate the GPU duty cycle metric
+showInstalling "The Prometheus Rule (custom GPU metrics rule)"
+kubectl apply -n prometheus -f prom-operator/stackdriver-exporter-gpu-rules.yaml

package/src/helm-charts/prom-operator/helmup.plugin

@@ -6,7 +6,7 @@ addHelmRepo prometheus-community https://prometheus-community.github.io/helm-cha
 NS="--namespace prometheus"
 
 showInstalling "The Prometheus Operator (kube-prometheus-stack)"
-[ -z "$PROMETHEUS_STACK_CHART_VERSION" ] && PROMETHEUS_STACK_CHART_VERSION="74"
+[ -z "$PROMETHEUS_STACK_CHART_VERSION" ] && PROMETHEUS_STACK_CHART_VERSION="75"
 
 # https://artifacthub.io/packages/helm/prometheus-community/kube-prometheus-stack
 helm upgrade $NS --install prometheus-stack prometheus-community/kube-prometheus-stack \
@@ -27,15 +27,4 @@ helm upgrade $NS --install stackdriver-exporter prometheus-community/prometheus-
                  --values prom-operator/stackdriver-exporter.yaml \
                  --version $STACKDRIVER_EXPORTER_CHART_VERSION $HELM_WHAT
 
-# https://artifacthub.io/packages/helm/prometheus-community/prometheus-adapter
-showInstalling "The Prometheus Adapter (prom-operator custom HPA metrics)"
-[ -z "$PROMETHEUS_ADAPTER_CHART_VERSION" ] && PROMETHEUS_ADAPTER_CHART_VERSION="4"
-helm upgrade $NS --install prometheus-adapter prometheus-community/prometheus-adapter \
-                 --values prom-operator/prometheus-adapter.yaml \
-                 --version $PROMETHEUS_ADAPTER_CHART_VERSION $HELM_WHAT
-
-# finally, apply the PrometheusRule to generate the GPU duty cycle metric
-showInstalling "The Prometheus Rule (custom GPU metrics rule)"
-kubectl apply -n prometheus -f prom-operator/stackdriver-exporter-gpu-rules.yaml
-
 removeHelmRepo prometheus-community

package/src/helm-charts/prom-operator/rules/elasticsearch-rules.yaml

@@ -174,7 +174,14 @@ spec:
 #               VALUE = {{ $value }}
 #               LABELS = {{ $labels }}
         - alert: ElasticsearchHighIndexingLatency
-          expr: elasticsearch_indices_indexing_index_time_seconds_total / elasticsearch_indices_indexing_index_total > 0.010 # awesome was 0.0005
+          expr: | # ChatGPT rewrite of awesome rule due to noise
+            (
+              rate(elasticsearch_indices_indexing_index_time_seconds_total[5m])
+              /
+              rate(elasticsearch_indices_indexing_index_total[5m])
+            ) > 0.02
+            and
+            rate(elasticsearch_indices_indexing_index_total[5m]) > 0.5
           for: 10m
           labels:
             severity: warning

package/src/helm-charts/prom-operator/rules/prometheus-rules.yaml

@@ -218,8 +218,8 @@ spec:
                 VALUE = {{ $value }}
                 LABELS = {{ $labels }}
         - alert: PrometheusTargetScrapeDuplicate
-          expr: increase(prometheus_target_scrapes_sample_duplicate_timestamp_total[5m]) > 0
-          for: 0m
+          expr: increase(prometheus_target_scrapes_sample_duplicate_timestamp_total[5m]) > 5
+          for: 5m # awesome was => 0m
           labels:
             severity: warning
           annotations:

package/src/helm-charts/resource-server/values-local.yaml

@@ -4,7 +4,7 @@
 #     - npm run sequelizeRunMigrations
 #
 image:
-  tag: # v1.14.1
+  tag: # v1.16.0
 
 config:
   LOG_CONFIG: '{"type":"pino","level":"warn"}'

package/src/helmup.sh

@@ -50,7 +50,6 @@ AUXILIARY=(
 
 # zombie-killer re-enabled 20240801
 SYSTEM=(
-  prom-operator
   elasticsearch8
   redis
   cnpg-operator
@@ -62,6 +61,7 @@ SYSTEM=(
   velero
   cnpg-db-psql-stack
   cnpg-db-tsdb-basic
+  prom-operator
 )
 
 SYSTEM_LEGACY=(

package/.vscode/settings.json

@@ -1,3 +0,0 @@
-{
-    "editor.formatOnSave": false
-}