DOP-4127: Monorepo builds without Makefiles

.vscode/launch.json

@@ -5,13 +5,15 @@
   "version": "0.2.0",
   "configurations": [
     {
+      "name": "Docker: Attach to Autobuilder",
       "type": "node",
-      "request": "launch",
-      "name": "Launch Program",
-      "skipFiles": ["<node_internals>/**"],
-      "program": "${workspaceFolder}/modules/oas-page-builder/tests/unit/services/pageBuilder.test.ts",
-      "preLaunchTask": "tsc: build - tsconfig.json",
-      "outFiles": ["${workspaceFolder}/build/**/*.js"]
+      "request": "attach",
+      "port": 9229,
+      "address": "localhost",
+      "localRoot": "${workspaceFolder}",
+      "remoteRoot": "/home/docsworker-xlarge",
+      "autoAttachChildProcesses": false,
+      "protocol": "inspector"
     }
   ]
 }

.vscode/settings.json

@@ -0,0 +1,3 @@
+{
+  "debug.javascript.autoAttachFilter": "disabled"
+}

Dockerfile.local

@@ -1,4 +1,4 @@
-FROM arm64v8/ubuntu:20.04
+FROM arm64v8/ubuntu:20.04 as initial
 ARG NPM_BASE_64_AUTH
 ARG NPM_EMAIL
 ARG SNOOTY_PARSER_VERSION=0.15.0
@@ -7,10 +7,13 @@ ARG MUT_VERSION=0.10.7
 ARG REDOC_CLI_VERSION=1.2.3
 ARG NPM_BASE_64_AUTH
 ARG NPM_EMAIL
+ARG WORK_DIRECTORY=/home/docsworker-xlarge
+
 ENV DEBIAN_FRONTEND=noninteractive
+WORKDIR ${WORK_DIRECTORY}
 
 # helper libraries for docs builds
-RUN apt-get update && apt-get install -y vim git unzip zip
+RUN apt-get update && apt-get install -y vim git unzip zip chromium-browser rsync
 
 # get node 18
 # https://gist.github.com/RinatMullayanov/89687a102e696b1d4cab
@@ -20,7 +23,6 @@ RUN apt-get install --yes nodejs
 RUN apt-get install --yes build-essential
 RUN apt-get install --yes python3-pip libxml2-dev libxslt-dev python-dev pkg-config
 
-WORKDIR /app
 
 RUN python3 -m pip install poetry
 
@@ -39,53 +41,66 @@ RUN git clone -b v${MUT_VERSION} --depth 1 https://github.com/mongodb/mut.git \
     && make package \
     && mv dist/mut /opt/
 
-RUN curl  -L -o redoc.zip https://github.com/mongodb-forks/redoc/archive/refs/tags/v${REDOC_CLI_VERSION}.zip \
-    && unzip redoc.zip \
-    && mv redoc-${REDOC_CLI_VERSION} redoc/ 
-
-ENV PATH="${PATH}:/opt/snooty:/opt/mut:/app/.local/bin"
+ENV PATH="${PATH}:/opt/snooty:/opt/mut:/${WORK_DIRECTORY}/.local/bin"
 
 # setup user and root directory
-RUN useradd -ms /bin/bash docsworker
-RUN chmod 755 -R /app
-RUN chown -Rv docsworker /app
-USER docsworker
+RUN useradd -ms /bin/bash docsworker-xlarge
+RUN chmod 755 -R ${WORK_DIRECTORY}
+RUN chown -Rv docsworker-xlarge ${WORK_DIRECTORY}
+USER docsworker-xlarge
 
 # install snooty frontend and docs-tools
 RUN git clone -b v${SNOOTY_FRONTEND_VERSION} --depth 1 https://github.com/mongodb/snooty.git       \
     && cd snooty                                                                                   \
     && npm ci --legacy-peer-deps --omit=dev                                                        
 
+RUN curl https://raw.githubusercontent.com/mongodb/docs-worker-pool/meta/makefiles/shared.mk -o shared.mk
+
+
+RUN git clone -b @dop/redoc-cli@${REDOC_CLI_VERSION} --depth 1 https://github.com/mongodb-forks/redoc.git redoc \
+    # Install dependencies for Redoc CLI
+    && cd redoc/ \
+    && npm ci --prefix cli/ --omit=dev
+
+FROM initial as persistence
+
 RUN mkdir -p modules/persistence && chmod 755 modules/persistence
 COPY modules/persistence/package*.json ./modules/persistence/
 RUN cd ./modules/persistence \
     && npm ci --legacy-peer-deps 
+# Build persistence module
+
+COPY --chown=docsworker-xlarge modules/persistence/tsconfig*.json ./modules/persistence
+COPY --chown=docsworker-xlarge modules/persistence/src ./modules/persistence/src/
+COPY --chown=docsworker-xlarge modules/persistence/index.ts ./modules/persistence
+
+RUN cd ./modules/persistence \
+    && npm run build:esbuild
+
+FROM initial as oas
 
 RUN  mkdir -p modules/oas-page-builder && chmod 755 modules/oas-page-builder
 COPY modules/oas-page-builder/package*.json ./modules/oas-page-builder/
 RUN cd ./modules/oas-page-builder \
     && npm ci --legacy-peer-deps 
+# Build modules
+# OAS Page Builder
+COPY --chown=docsworker-xlarge modules/oas-page-builder/tsconfig*.json ./modules/oas-page-builder
+COPY --chown=docsworker-xlarge modules/oas-page-builder/src ./modules/oas-page-builder/src/
+COPY --chown=docsworker-xlarge modules/oas-page-builder/index.ts ./modules/oas-page-builder
 
-# Root project build
-COPY package*.json ./
-RUN npm ci --legacy-peer-deps
-# Build persistence module
+RUN cd ./modules/oas-page-builder \
+    && npm run build:esbuild
 
-COPY --chown=docsworker modules/persistence/tsconfig*.json ./modules/persistence
-COPY --chown=docsworker modules/persistence/src ./modules/persistence/src/
-COPY --chown=docsworker modules/persistence/index.ts ./modules/persistence
+FROM initial as root
 
-RUN cd ./modules/persistence \
-    && npm run build
+COPY --from=persistence --chown=docsworker-xlarge ${WORK_DIRECTORY}/modules/persistence/dist/ ./modules/persistence
+COPY --from=oas --chown=docsworker-xlarge ${WORK_DIRECTORY}/modules/oas-page-builder/dist/ ./modules/oas-page-builder
 
-# Build modules
-# OAS Page Builder
-COPY --chown=docsworker modules/oas-page-builder/tsconfig*.json ./modules/oas-page-builder
-COPY --chown=docsworker modules/oas-page-builder/src ./modules/oas-page-builder/src/
-COPY --chown=docsworker modules/oas-page-builder/index.ts ./modules/oas-page-builder
+# Root project build
+COPY package*.json ./
+RUN npm ci --legacy-peer-deps
 
-RUN cd ./modules/oas-page-builder \
-    && npm run build
 
 COPY tsconfig*.json ./
 COPY config config/
@@ -94,8 +109,13 @@ COPY src src/
 
 RUN npm run build:esbuild
 
+ENV PERSISTENCE_MODULE_PATH=${WORK_DIRECTORY}/modules/persistence/index.js
+ENV OAS_MODULE_PATH=${WORK_DIRECTORY}/modules/oas-page-builder/index.js
+ENV REDOC_PATH=${WORK_DIRECTORY}/redoc/cli/index.js
+
+RUN mkdir -p modules/persistence && chmod 755 modules/persistence
 RUN mkdir repos && chmod 755 repos
 
 EXPOSE 3000
 
-CMD ["node", "--enable-source-maps", "dist/entrypoints/localApp.js"]
+CMD ["node", "--inspect-brk=0.0.0.0", "--enable-source-maps", "dist/entrypoints/onDemandApp.js"]

README.md

@@ -21,49 +21,81 @@ To add a new property:
 - Go to `infrastructure/ecs-main/ecs-service.yml` `TaskDefinition` section
 - Add the new property to the `ContainerDefinitions`/`Environment` section
 
-## Build and Run Docker Image for local testing
+## Debug the Autobuilder for Local Testing
 
-The npm build args are required for the portion of the dockerfile that installs the [snooty-frontend]. `NPM_CONFIG__AUTH`
-and `NPM_CONFIG_EMAIL` are environment variables available in our working directory. `NPM_CONFIG_{OPTION}` environment
-variables can actually be used instead of the `~/.npmrc` file. The reason we need the build args to be `NPM_BASE_64_AUTH`
-and `NPM_EMAIL` is because that's what's expected in the `.npmrc` within [snooty-frontend].
+### Setup
 
-```shell
-docker build --tag=workerpool --build-arg NPM_BASE_64_AUTH=${NPM_CONFIG__AUTH} --build-arg NPM_EMAIL=${NPM_CONFIG_EMAIL} .
-```
+To debug the Autobuilder for local testing, you first need to ensure the following has been done:
 
-```shell
-docker run \
-	--env MONGO_ATLAS_USERNAME \
-	--env MONGO_ATLAS_PASSWORD \
-	--env AWS_ACCESS_KEY_ID \
-	--env AWS_SECRET_ACCESS_KEY \
-	--env GITHUB_BOT_USERNAME \
-	--env GITHUB_BOT_PASSWORD \
-	--env DB_NAME \
-	--env XLARGE \
-	--env SNOOTY_ENV \
-	--env FASTLY_TOKEN \
-	--env FASTLY_DOCHUB_MAP \
-	--env FASTLY_SERVICE_ID \
-	workerpool
-```
+1. Docker is running
+2. The `~/.aws/credentials` file contains unexpired credentials for the `default` profile
 
-- `MONGO_ATLAS_USERNAME` and `MONGO_ATLAS_PASSWORD` is username/password of atlas database
-- `AWS_ACCESS_KEY_ID` and `AWS_SECRET_ACCESS_KEY` are needed for uploading to S3 via [mut](https://github.com/mongodb/mut)
-- `GITHUB_BOT_USERNAME` and `GITHUB_BOT_PASSWORD` are needed so builder can access private repos
-- `DB_NAME` allows the indication of a pool database (pool, pool_test)
-- `XLARGE` true or false indicates whether this instance will run on an XLARGE server or not
-- `SNOOTY_ENV` indicates whether the build environment is stage, prod, or dev
-- `FASTLY_TOKEN` is needed for connecting to the Fastly edge dictionary
-- `FASTLY_DOCHUB_MAP` is the id of the redirect map that we publish dochub links to
-- `FASTLY_SERVICE_ID` is the id of the service used for dochub
+For retrieving credentials, head to AWS and under `Docs Platform`, click on `Command line or programmatic access`.
+![AWS console](images/aws-console-admin.png)
 
-If you are running a local version of the docker image for testing, we have a separate staging environment setup. Testing in this environment is automated through the "stage" branch. Add the following env variables to the `docker run` command:
+Copy the value in option 2, `Manually add a profile to your AWS credentials file (Short-term credentials)`.
 
+![Alt text](images/aws-credentials.png)
+
+From there, paste this value in `~/.aws/credentials`, and replace the randomly generated profile (which looks something like `[123456789_AdministratorAccess]`) with `[default]`.
+You should now have the correct credentials to run the debugger.
+
+_**NOTE: credentials expire pretty quickly. Not sure how exactly how long they last for, but in my experience they expire in approximately 30 minutes.**_
+
+You should now be all set to run the debugger command:
+
+`npm run debug`
+
+To view all of the options for the command, you can run:
+
+`npm run debug -- --help`
+
+Here is an example of running the local debugger for `cloud-docs`:
+
+`npm run debug -- -o 10gen -n cloud-docs`
+
+Here is an example of running the local debugger for `docs-monorepo/docs-landing` on branch `groot`:
+
+`npm run debug -- -o 10gen -n docs-monorepo -d cloud-docs -b groot`
+
+By default, the environment that is used for the local Autobuilder is `stg`.
+
+### Debugger Behavior
+
+When the command is run, there are several steps that occur before the Autobuilder begins:
+
+1. Environment variables and other information are pulled from Parameter Store
+2. The GitHub repository is queried for data to create the job
+3. The container is built
+   - NOTE: If you have not run the debug command before, the build will take a substantial amount of time (approximately 10-15 minutes).
+     Subsequent builds will be much shorter, especially if the changes are just code changes. If just a code change is made after the initial build, it should only take a few seconds for the build to complete and the container to run. Changes such as updating the version of the Snooty Parser, or the Redoc CLI will cause the builds to take much longer, but these happen much less frequently. The majority of the build should be on the order of a few seconds.
+4. The data from step 2 is then added as a record in the `pool_test.queue`.
+5. The container is then run, and waits for the user to connect to it via the VSCode debugger.
+
+Once the container starts successfully, you should see something like the following message:
+
+`Debugger listening on ws://0.0.0.0:9229/....`
+
+To connect, click on the debug tab on the left side of your VSCode editor. Make sure the dropdown to the right of the green play button is set to the `Docker: Attach to Autobuilder` configuration. Press the green play button, and you will attach to the container.
+
+### Troubleshooting
+
+The most frequent cause of build failures will be related to expired AWS credentials, or not having Docker running. Also, if you haven't run `npm ci` in a while, you will need to do so as a new dependency was added to run the command.
+
+Occasionally, errors may occur inexplicably, and the error messages may seem unrelated to any change made. Oftentimes, running the following commands can resolve these sporadic issues:
+
+```sh
+docker image prune
+docker container prune
 ```
---env DB_NAME
-```
+
+Also, another potential error could be due to the Dockerfile.local not being updated. If you are not seeing changes that are occurring in the Autobuilder in another environment, this may be why. For example, the Dockerfile.local could be using an older version of the Snooty Parser.
+
+![Alt text](images/vsode-debugger.png)
+
+By default, the container will break at the first line of code, which will be in a file called `bind.js`. Press the fast-forward button to continue the execution. You are also able to add other breakpoints to stop the application. Once the application is complete, press `CTRL + C` for the terminal to exit out of the connection to the container.
+
+If you receive `CredentialsProviderError: Could not load credentials from any providers`, make sure that there is no env `AWS_PROFILE` defined as a different profile anywhere (such as in the global `~/.zshrc` file). Otherwise, ensure that `AWS_PROFILE` matches the same profile defined in `~/.aws/credentials`.
 
 ## Run Tests
 

api/controllers/v1/github.ts

@@ -6,9 +6,10 @@
 import { markBuildArtifactsForDeletion, validateJsonWebhook } from '../../handlers/github';
 import { DocsetsRepository } from '../../../src/repositories/docsetsRepository';
 import { ReposBranchesDocsetsDocument } from '../../../modules/persistence/src/services/metadata/repos_branches';
+import { PushEvent } from '@octokit/webhooks-types';
 
 async function prepGithubPushPayload(
   githubEvent: any,
   repoBranchesRepository: RepoBranchesRepository,
   prefix: string,
   repoInfo: ReposBranchesDocsetsDocument
@@ -60,7 +61,7 @@
   };
 }
 
 export const TriggerBuild = async (event: any = {}, context: any = {}): Promise<any> => {
   const client = new mongodb.MongoClient(c.get('dbUrl'));
   await client.connect();
   const db = client.db(c.get('dbName'));
@@ -77,7 +78,28 @@
       body: errMsg,
     };
   }
-  const body = JSON.parse(event.body);
+  if (!event.body) {
+    const err = 'Trigger build does not have a body in event payload';
+    consoleLogger.error('TriggerBuildError', err);
+    return {
+      statusCode: 400,
+      headers: { 'Content-Type': 'text/plain' },
+      body: err,
+    };
+  }
+
+  let body: PushEvent;
+  try {
+    body = JSON.parse(event.body) as PushEvent;
+  } catch (e) {
+    consoleLogger.error('[TriggerBuild]', `ERROR! Could not parse event.body ${e}`);
+    console.log(`event: ${event} and event body: ${event.body}`);
+    return {
+      statusCode: 502,
+      headers: { 'Content-Type': 'text/plain' },
+      body: ' ERROR! Could not parse event.body',
+    };
+  }
 
   if (body.deleted) {
     return {

api/controllers/v1/jobs.ts

@@ -13,7 +13,7 @@
 import { notifyBuildSummary, snootyBuildComplete } from '../../handlers/jobs';
 import { DocsetsRepository } from '../../../src/repositories/docsetsRepository';
 
 export const TriggerLocalBuild = async (event: any = {}, context: any = {}): Promise<any> => {
   const client = new mongodb.MongoClient(c.get('dbUrl'));
   await client.connect();
   const db = client.db(c.get('dbName'));
@@ -90,7 +90,7 @@
             await SubmitArchiveJob(jobId);
             break;
           default:
-            consoleLogger.error(jobId, 'Invalid status');
+            consoleLogger.error(jobId, `Invalid status: ${jobStatus}`);
             break;
         }
       } catch (err) {

api/controllers/v2/github.ts

@@ -18,7 +18,8 @@ async function prepGithubPushPayload(
   githubEvent: PushEvent,
   repoBranchesRepository: RepoBranchesRepository,
   prefix: string,
-  repoInfo: ReposBranchesDocsetsDocument
+  repoInfo: ReposBranchesDocsetsDocument,
+  directory?: string
 ): Promise<Omit<EnhancedJob, '_id'>> {
   const branch_name = githubEvent.ref.split('/')[2];
   const branch_info = await repoBranchesRepository.getRepoBranchAliases(
@@ -53,6 +54,7 @@ async function prepGithubPushPayload(
       urlSlug: urlSlug,
       prefix: prefix,
       project: project,
+      directory: directory,
     },
     logs: [],
   };
@@ -110,7 +112,7 @@ export const TriggerBuild = async (event: APIGatewayEvent): Promise<APIGatewayPr
   async function createAndInsertJob(path?: string) {
     const repoInfo = await docsetsRepository.getRepo(body.repository.name, path);
     const jobPrefix = repoInfo?.prefix ? repoInfo['prefix'][env] : '';
-    const job = await prepGithubPushPayload(body, repoBranchesRepository, jobPrefix, repoInfo);
+    const job = await prepGithubPushPayload(body, repoBranchesRepository, jobPrefix, repoInfo, path);
 
     consoleLogger.info(job.title, 'Creating Job');
     const jobId = await jobRepository.insertJob(job, c.get('jobsQueueUrl'));
@@ -128,10 +130,10 @@ export const TriggerBuild = async (event: APIGatewayEvent): Promise<APIGatewayPr
           ownerName: body.repository.owner.name,
           updatedFilePaths: getUpdatedFilePaths(body.head_commit),
         });
-        consoleLogger.info('monoRepoPaths', `Monorepo Paths with new changes: ${monorepoPaths}`);
+        consoleLogger.info(body.repository.full_name, `Monorepo Paths with new changes: ${monorepoPaths}`);
       }
     } catch (error) {
-      console.warn('Warning, attempting to get repo paths caused an error', error);
+      consoleLogger.warn('Warning, attempting to get monorepo paths caused an error', error);
     }
 
     /* Create and insert Job for each monorepo project that has changes */
@@ -141,7 +143,7 @@ export const TriggerBuild = async (event: APIGatewayEvent): Promise<APIGatewayPr
       if (path.split('/').length > 1) continue;
 
       try {
-        await createAndInsertJob(`/${path}`);
+        await createAndInsertJob(path);
       } catch (err) {
         return {
           statusCode: 500,

api/controllers/v2/jobs.ts

@@ -72,7 +72,7 @@ export const HandleJobs = async (event: SQSEvent): Promise<void> => {
             await notifyBuildSummary(jobId);
             break;
           default:
-            consoleLogger.error(jobId, 'Invalid status');
+            consoleLogger.error(jobId, `Invalid status: ${jobStatus}`);
             break;
         }
       } catch (err) {
@@ -150,7 +150,7 @@ async function NotifyBuildProgress(jobId: string): Promise<void> {
     return;
   }
 
-  const jobTitle = fullDocument.title;
+  const jobTitle = `${fullDocument.title}${fullDocument.payload.directory ? `/${fullDocument.payload.directory}` : ''}`;
   const username = fullDocument.user;
   const repoEntitlementRepository = new RepoEntitlementsRepository(db, c, consoleLogger);
   const entitlement = await repoEntitlementRepository.getSlackUserIdByGithubUsername(username);