[7.17] Bump postcss to ^8 (#136303)

package.json

@@ -648,7 +648,7 @@
     "antlr4ts-cli": "^0.5.0-alpha.3",
     "apidoc-markdown": "^7.2.4",
     "argsplit": "^1.0.5",
-    "autoprefixer": "^9.7.4",
+    "autoprefixer": "^10.4.7",
     "axe-core": "^4.0.2",
     "babel-jest": "^27.5.1",
     "babel-loader": "^8.2.2",
@@ -719,7 +719,7 @@
     "gulp-babel": "^8.0.0",
     "gulp-brotli": "^3.0.0",
     "gulp-gzip": "^1.4.2",
-    "gulp-postcss": "^8.0.0",
+    "gulp-postcss": "^9.0.1",
     "gulp-sourcemaps": "2.6.5",
     "gulp-terser": "^2.0.1",
     "gulp-zip": "^5.0.2",
@@ -775,11 +775,13 @@
     "pirates": "^4.0.1",
     "pixelmatch": "^5.3.0",
     "playwright": "^1.17.1",
-    "postcss": "^7.0.32",
-    "postcss-loader": "^3.0.0",
-    "postcss-prefix-selector": "^1.7.2",
-    "prettier": "^2.4.0",
-    "pretty-ms": "5.0.0",
+    "pngjs": "^3.4.0",
+    "postcss": "^8.4.14",
+    "postcss-loader": "^4.2.0",
+    "postcss-prefix-selector": "^1.16.0",
+    "prettier": "^2.7.1",
+    "pretty-format": "^27.5.1",
+    "pretty-ms": "^5.0.0",
     "proxy": "^1.0.2",
     "q": "^1.5.1",
     "react-test-renderer": "^16.12.0",

packages/kbn-optimizer/src/worker/webpack.config.ts

@@ -154,8 +154,8 @@ export function getWebpackConfig(bundle: Bundle, bundleRefs: BundleRefs, worker:
                   loader: 'postcss-loader',
                   options: {
                     sourceMap: !worker.dist,
-                    config: {
-                      path: require.resolve('@kbn/optimizer/postcss.config.js'),
+                    postcssOptions: {
+                      config: require.resolve('@kbn/optimizer/postcss.config.js'),
                     },
                   },
                 },

packages/kbn-storybook/src/webpack.config.ts

@@ -93,8 +93,8 @@ export default function ({ config: storybookConfig }: { config: Configuration })
             {
               loader: 'postcss-loader',
               options: {
-                config: {
-                  path: require.resolve('@kbn/optimizer/postcss.config.js'),
+                postcssOptions: {
+                  config: require.resolve('@kbn/optimizer/postcss.config.js'),
                 },
               },
             },

src/core/server/saved_objects/migrationsv2/actions/create_index.ts

@@ -55,72 +55,74 @@ export const createIndex = ({
   mappings,
   aliases = [],
 }: CreateIndexParams): TaskEither.TaskEither<RetryableEsClientError, 'create_index_succeeded'> => {
-  const createIndexTask: TaskEither.TaskEither<RetryableEsClientError, AcknowledgeResponse> =
-    () => {
-      const aliasesObject = aliasArrayToRecord(aliases);
+  const createIndexTask: TaskEither.TaskEither<
+    RetryableEsClientError,
+    AcknowledgeResponse
+  > = () => {
+    const aliasesObject = aliasArrayToRecord(aliases);
 
-      return client.indices
-        .create(
-          {
-            index: indexName,
-            // wait until all shards are available before creating the index
-            // (since number_of_shards=1 this does not have any effect atm)
-            wait_for_active_shards: WAIT_FOR_ALL_SHARDS_TO_BE_ACTIVE,
-            // Wait up to 60s for the cluster state to update and all shards to be
-            // started
-            timeout: DEFAULT_TIMEOUT,
-            body: {
-              mappings,
-              aliases: aliasesObject,
-              settings: {
-                index: {
-                  // ES rule of thumb: shards should be several GB to 10's of GB, so
-                  // Kibana is unlikely to cross that limit.
-                  number_of_shards: 1,
-                  auto_expand_replicas: INDEX_AUTO_EXPAND_REPLICAS,
-                  // Set an explicit refresh interval so that we don't inherit the
-                  // value from incorrectly configured index templates (not required
-                  // after we adopt system indices)
-                  refresh_interval: '1s',
-                  // Bump priority so that recovery happens before newer indices
-                  priority: 10,
-                },
+    return client.indices
+      .create(
+        {
+          index: indexName,
+          // wait until all shards are available before creating the index
+          // (since number_of_shards=1 this does not have any effect atm)
+          wait_for_active_shards: WAIT_FOR_ALL_SHARDS_TO_BE_ACTIVE,
+          // Wait up to 60s for the cluster state to update and all shards to be
+          // started
+          timeout: DEFAULT_TIMEOUT,
+          body: {
+            mappings,
+            aliases: aliasesObject,
+            settings: {
+              index: {
+                // ES rule of thumb: shards should be several GB to 10's of GB, so
+                // Kibana is unlikely to cross that limit.
+                number_of_shards: 1,
+                auto_expand_replicas: INDEX_AUTO_EXPAND_REPLICAS,
+                // Set an explicit refresh interval so that we don't inherit the
+                // value from incorrectly configured index templates (not required
+                // after we adopt system indices)
+                refresh_interval: '1s',
+                // Bump priority so that recovery happens before newer indices
+                priority: 10,
               },
             },
           },
-          { maxRetries: 0 /** handle retry ourselves for now */ }
-        )
-        .then((res) => {
+        },
+        { maxRetries: 0 /** handle retry ourselves for now */ }
+      )
+      .then((res) => {
+        /**
+         * - acknowledged=false, we timed out before the cluster state was
+         *   updated on all nodes with the newly created index, but it
+         *   probably will be created sometime soon.
+         * - shards_acknowledged=false, we timed out before all shards were
+         *   started
+         * - acknowledged=true, shards_acknowledged=true, index creation complete
+         */
+        return Either.right({
+          acknowledged: Boolean(res.body.acknowledged),
+          shardsAcknowledged: res.body.shards_acknowledged,
+        });
+      })
+      .catch((error) => {
+        if (error?.body?.error?.type === 'resource_already_exists_exception') {
           /**
-           * - acknowledged=false, we timed out before the cluster state was
-           *   updated on all nodes with the newly created index, but it
-           *   probably will be created sometime soon.
-           * - shards_acknowledged=false, we timed out before all shards were
-           *   started
-           * - acknowledged=true, shards_acknowledged=true, index creation complete
+           * If the target index already exists it means a previous create
+           * operation had already been started. However, we can't be sure
+           * that all shards were started so return shardsAcknowledged: false
            */
           return Either.right({
-            acknowledged: Boolean(res.body.acknowledged),
-            shardsAcknowledged: res.body.shards_acknowledged,
+            acknowledged: true,
+            shardsAcknowledged: false,
           });
-        })
-        .catch((error) => {
-          if (error?.body?.error?.type === 'resource_already_exists_exception') {
-            /**
-             * If the target index already exists it means a previous create
-             * operation had already been started. However, we can't be sure
-             * that all shards were started so return shardsAcknowledged: false
-             */
-            return Either.right({
-              acknowledged: true,
-              shardsAcknowledged: false,
-            });
-          } else {
-            throw error;
-          }
-        })
-        .catch(catchRetryableEsClientErrors);
-    };
+        } else {
+          throw error;
+        }
+      })
+      .catch(catchRetryableEsClientErrors);
+  };
 
   return pipe(
     createIndexTask,

src/core/server/saved_objects/migrationsv2/actions/update_and_pickup_mappings.ts

@@ -41,33 +41,35 @@ export const updateAndPickupMappings = ({
   RetryableEsClientError,
   UpdateAndPickupMappingsResponse
 > => {
-  const putMappingTask: TaskEither.TaskEither<RetryableEsClientError, 'update_mappings_succeeded'> =
-    () => {
-      return client.indices
-        .putMapping({
-          index,
-          timeout: DEFAULT_TIMEOUT,
-          body: mappings,
-        })
-        .then((res) => {
-          // Ignore `acknowledged: false`. When the coordinating node accepts
-          // the new cluster state update but not all nodes have applied the
-          // update within the timeout `acknowledged` will be false. However,
-          // retrying this update will always immediately result in `acknowledged:
-          // true` even if there are still nodes which are falling behind with
-          // cluster state updates.
-          // For updateAndPickupMappings this means that there is the potential
-          // that some existing document's fields won't be picked up if the node
-          // on which the Kibana shard is running has fallen behind with cluster
-          // state updates and the mapping update wasn't applied before we run
-          // `pickupUpdatedMappings`. ES tries to limit this risk by blocking
-          // index operations (including update_by_query used by
-          // updateAndPickupMappings) if there are pending mappings changes. But
-          // not all mapping changes will prevent this.
-          return Either.right('update_mappings_succeeded' as const);
-        })
-        .catch(catchRetryableEsClientErrors);
-    };
+  const putMappingTask: TaskEither.TaskEither<
+    RetryableEsClientError,
+    'update_mappings_succeeded'
+  > = () => {
+    return client.indices
+      .putMapping({
+        index,
+        timeout: DEFAULT_TIMEOUT,
+        body: mappings,
+      })
+      .then((res) => {
+        // Ignore `acknowledged: false`. When the coordinating node accepts
+        // the new cluster state update but not all nodes have applied the
+        // update within the timeout `acknowledged` will be false. However,
+        // retrying this update will always immediately result in `acknowledged:
+        // true` even if there are still nodes which are falling behind with
+        // cluster state updates.
+        // For updateAndPickupMappings this means that there is the potential
+        // that some existing document's fields won't be picked up if the node
+        // on which the Kibana shard is running has fallen behind with cluster
+        // state updates and the mapping update wasn't applied before we run
+        // `pickupUpdatedMappings`. ES tries to limit this risk by blocking
+        // index operations (including update_by_query used by
+        // updateAndPickupMappings) if there are pending mappings changes. But
+        // not all mapping changes will prevent this.
+        return Either.right('update_mappings_succeeded' as const);
+      })
+      .catch(catchRetryableEsClientErrors);
+  };
 
   return pipe(
     putMappingTask,