Clean up leftover bigquery datasets

PR-URL: https://github.com/hasura/graphql-engine-mono/pull/8543
GitOrigin-RevId: 9463e50aa4bf62e12e39b4bebdf551e11c824897

scripts/cleanup-bigquery.js

@@ -0,0 +1,44 @@
+// Delete a bunch of bigquery test datasets.
+//
+// Note that we can only select the first 50 datasets at a time, so you might
+// have to run this a few times to clear it out completely.
+
+const { exec } = require('child_process')
+const util = require('util')
+
+const allDatasetsCommand = "bq ls --project_id regency-polecat-beehive --format=prettyjson | jq '.[]  | .datasetReference.datasetId | select(startswith(\"hasura_test_\"))'"
+const singleDatasetCommand = dataset => `bq show --format=prettyjson regency-polecat-beehive:${dataset}`
+const removeDatasetCommand = dataset => `bq rm --dataset=true --force=true --recursive=true regency-polecat-beehive:${dataset}`
+
+const now = Date.now()
+
+const removeDataset = dataset =>
+  util.promisify(exec)(removeDatasetCommand(dataset))
+    .then(_ => console.log('Deleted ' + dataset))
+
+const checkDataset = dataset =>
+  util.promisify(exec)(singleDatasetCommand(dataset))
+    .then(({ stdout }) => {
+      const parsed = JSON.parse(stdout)
+            lastUpdate = parsed.lastModifiedTime
+
+      if (now - (1000 * 60 * 60 * 6) > lastUpdate) {
+        console.log('Deleting ' + dataset)
+        removeDataset(dataset)
+      } else {
+        console.log('Not deleting ' + dataset)
+      }
+    })
+    .catch(_ => console.log('Skipping ' + dataset))
+
+util.promisify(exec)(allDatasetsCommand)
+  .then(({ stdout }) => {
+    const datasets =
+            stdout
+              .split('\n')
+              .filter(x => x != '')
+              .map(x => x.substring(1, x.length - 1))
+
+    console.log('Found ' + datasets.length + ' datasets')
+    datasets.map(checkDataset)
+  })

scripts/cleanup-bigquery.sh

@@ -0,0 +1,13 @@
+#!/usr/bin/env bash
+
+# Clean up old BigQuery test datasets.
+
+# If we end the bigquery API test suite abruptly, the created datasets won't be
+# cleaned up. In which case, we end up with a bunch of `hasura_test_*` datasets
+# that hang around forever. When we get too many of these, we can run this
+# script, and it will delete them in batches. See the JavaScript file for more
+# information.
+
+gcloud auth login
+gcloud config set project regency-polecat-beehive
+node $(dirname "$0")/cleanup-bigquery.js