Conversation
<!-- Please complete the following sections as necessary. --> ### Description Creates a docker compose file that defines a local postgres database to use for development and testing. Also updates the README with instructions for how to run the local db. ### Ticket Resolves [Ticket #424](newjersey/innovation-platform-pm#424)
<!-- Please complete the following sections as necessary. --> ### Description Creates the `Feedback` Prisma data model, as well as the initial migration to load this schema into our postgres database. ### Ticket Resolves [Ticket #430](newjersey/innovation-platform-pm#430) (Define a basic schema for the feedback database using Prisma/Prisma Migrate) ### Approach - Added the `@prisma/client` package - Added `DATABASE_URL` to .env - This is set to the local db URL (`postgresql://postgres:postgres@localhost:5432/postgres`) - I've also updated the Bitwarden with the new .env - I changed the database server time zone from `UTC` to `US/Eastern` by editing the `TZ` env var in the `docker-compose.yml` - We originally chose UTC because we didn't want to manage between switching from EDT/EST, but the`US/Eastern` time zone automatically handles this for us. - With the current database schema, when we create a new feedback record, the `createdAt` column defaults to the time the record was created. By setting the database time zone to `US/Eastern`, it will automatically generate the timestamp in Eastern time. - So when we're looking up records in the database, the dates will be more intuitive since they will match NJ's time zone, which is generally the timezone that people are submitting feedback from. - This might also make integrating with Looker easier as we don't have to convert the time from another timezone when displaying dates. ### Steps to Test #### Loading the schema into the local db If this is your first time running the local db, follow the [instructions in the README](https://github.com/newjersey/feedback-api/blob/dev/README.md#running-the-local-dev-database) first. 1. Run `npm install` 2. Delete the `feedback-api_postgres-data` volume since we don't want old persisted data from the local db - Run `docker volume ls` to see a list of volumes - If the volume is present, run `docker volume rm feedback-api_postgres-data` to delete it 3. Start the local db by running `colima start` then `docker compose up -d` 4. Run `npx prisma migrate dev` to apply the migrations in `prisma/migrations/` to the local db 5. Generate the Prisma Client containing our database schema's data models with `npx prisma generate` #### Manual testing strategy I didn't want to open the integration testing can of works yet, I decided to manually test that inserting into the Feedback table with Prisma works using a ts script, specifically paying attention that the `createdAt` column is correct. 1. Installed ts-node without saving it to package.json by running `npm install ts-node --no-save` 2. Created a `test.ts` file in the `prisma/` directory with the following code: ```ts import { PrismaClient } from '..//generated/prisma/client'; const prisma = new PrismaClient(); async function main() { await prisma.feedback.deleteMany({}); const feedbackWithoutTime = await prisma.feedback.create({ data: { rating: true, pageUrl: 'testUrl', createdAtFromApi: new Date() } }); const feedbackWithTime = await prisma.feedback.create({ data: { rating: true, pageUrl: 'testUrl', createdAt: new Date(1671717203269), // Thursday, December 22, 2022 1:53:23.269 PM UTC createdAtFromApi: new Date() } }); console.log({ feedbackWithoutTime, feedbackWithTime }); } main() .then(async () => { await prisma.$disconnect(); }) .catch(async (e) => { console.error(e); await prisma.$disconnect(); process.exit(1); }); ``` 3. Ran the script with `npx ts-node prisma/test.ts`. This produced the following console output: ```bash { feedbackWithoutTime: { id: 1, createdAt: 2025-07-22T19:43:28.141Z, createdAtFromApi: 2025-07-22T19:43:28.139Z, rating: true, pageUrl: 'testUrl', comment: null, email: null }, feedbackWithTime: { id: 2, createdAt: 2022-12-22T13:53:23.269Z, createdAtFromApi: 2025-07-22T19:43:28.147Z, rating: true, pageUrl: 'testUrl', comment: null, email: null } } ``` - Success criteria - The ids are being generated sequentially - For `feedbackWithoutTime`, the createdAt matches the time at which I ran the script (formatted in UTC) - For `feedbackWith`, the createdAt matches the inputted DateTime object (formatted in UTC) 4. Then, checked that the objects were created as expected in the database itself. Ran `psql postgresql://postgres:postgres@localhost:5432/postgres` to start the psql console. 5. Within the psql console, ran `SELECT * FROM feedback;`. This gave the following output ``` id | created_at | created_at_from_api | rating | page_url | comment | email ----+----------------------------+----------------------------+--------+----------+---------+------- 1 | 2025-07-22 15:43:28.141-04 | 2025-07-22 15:43:28.139-04 | t | testUrl | | 2 | 2022-12-22 08:53:23.269-05 | 2025-07-22 15:43:28.147-04 | t | testUrl | | (2 rows) ``` - Success criteria - The `id`, `rating`, and `pageUrl` fields match what we inserted with prisma. - The `createdAt` and `createdAtFromApi` fields are displayed in "US/Eastern" time, i.e. it automatically selects the appropriate UTC offset depending on whether DST was in effect on the given date. - The Prisma camelCase fields (e.g. `createdAt`) are mapped to snake_case based on the `@map` attributes in prisma.schema.
<!-- Please complete the following sections as necessary. --> ### Description <!-- Summary of the changes, related issue, relevant motivation, and context --> This PR resolves [Configure Aurora Serverless v2 PostgreSQL](newjersey/innovation-platform-pm#464) by implementing Infrastructure as Code (IaC) for an AWS Aurora Serverless v2 PostgreSQL database. ### Approach <!-- Any changed dependencies, e.g. requires an install/update/migration, etc. --> **Using AWS Cloud Development Kit (CDK) and TypeScript:** - Referenced the existing dev VPC (will update to prod VPC once PR is approved and production deployment schedule is set) - Created supporting resources: subnet group, security group, and Aurora cluster - Exported VPC and security group for use in API stack (lambdas) - Updated `FeedbackApiStack` to accept shared `vpc` and `securityGroup` from `FeedbackDbStack` - Wrote tests to validate mocked VPC, creation of resources, and configuration settings - Used `cdk synth` to generate and validate CloudFormation output #### AWS Resource Documentation - [Work with the AWS CDK in TypeScript](https://docs.aws.amazon.com/cdk/v2/guide/work-with-cdk-typescript.html#typescript-newproject) - [How to create a CDK app](https://docs.aws.amazon.com/cdk/v2/guide/apps.html) - [Define an AWS Virtual Private Cloud](https://docs.aws.amazon.com/cdk/api/v2/docs/aws-cdk-lib.aws_ec2.Vpc.html) - [Create an RDS DB subnet group](https://docs.aws.amazon.com/cdk/api/v2/docs/aws-cdk-lib.aws_rds.SubnetGroup.html) - [Create an Amazon EC2 security group within a VPC](https://docs.aws.amazon.com/cdk/api/v2/docs/aws-cdk-lib.aws_ec2.SecurityGroup.html) - [Create a clustered database](https://docs.aws.amazon.com/cdk/api/v2/docs/aws-cdk-lib.aws_rds.DatabaseCluster.html) - [AWS CDK Test Assertions](https://docs.aws.amazon.com/cdk/api/v2/docs/aws-cdk-lib.assertions-readme.html) ### Steps to Test <!-- If this work affects a user's experience, provide steps to test these changes in-app. --> 1. Run `npm install` to install dependencies. 2. Run `npm run synth:dev` to synthesize the CDK app into a dev CloudFormation template. 3. 2. Run `npm run deploy-db:dev` to deploy the dev database AWS. 4. Verify synthesized template includes VPC, subnet group, security group, and Aurora cluster. #### **To check deployment (in AWS Console):** 1. Search for/Go to CloudFormation. 2. Choose your AWS region (e.g., us-east-1) at the top-right. 3. Look for a stack named `FeedbackDbStack`. 4. Confirm its status is ✅ CREATE_COMPLETE. ### Notes <!-- Additional information, key learnings, and future development considerations. --> - [x] Deployment requires AWS account authorization due to costs; awaiting leadership approval - [ ] Once PR is approved, we can: - [x] deploy to the dev account - [ ] merge PR into dev
<!-- Please complete the following sections as necessary. --> ### Description <!-- Summary of the changes, related issue, relevant motivation, and context --> Implements [#578 Set up alerts for new Feedback API (in Innov-Platform-Prod AWS account) ](newjersey/innovation-platform-pm#578). #### Context As part of migrating the Feedback API resources from the `Innov-RES-Dev` to the Platforms AWS accounts, we realized that our alarm for API errors was not defined in CDK (nor Serverless before we migrated to CDK). Previously in the `Innov-RES-Dev` account, we had an alarm named "Feedback API - Comments - Error" which tracked a metric by the same name. With @sannidhishukla's help we tracked this metric back down a metric filter on the [/aws/lambda/feedback-api-dev-comment log group](https://us-east-1.console.aws.amazon.com/cloudwatch/home?region=us-east-1#logsV2:log-groups/log-group/$252Faws$252Flambda$252Ffeedback-api-dev-comment$23metric-filters) log group which matches when the string "Error:" appears in logs within that log stream. For the sake of simplicity, instead of copying the previous approach, I decided to set up two new alarms that instead track two built-in metrics, `5XXError` and `4XXError`, on the API gateway. An additional benefit is that the alarm will trigger on all of our endpoints rather than just the comment endpoint. ### Approach <!-- Any changed dependencies, e.g. requires an install/update/migration, etc. --> **Using AWS Cloud Development Kit (CDK) and TypeScript:** - Defined two alarm constructs for the 5XXError and 4XXError alarms and added them to the existing FeedbackApiStack. - Created an SNS topic and added a subscription to send alerts to our #platform-eng-alerts Slack channel. #### AWS Resource Documentation - [Set a CloudWatch alarm](https://docs.aws.amazon.com/cdk/v2/guide/how-to-set-cw-alarm.html) - [Alarm construct](https://docs.aws.amazon.com/cdk/api/v2/docs/aws-cdk-lib.aws_cloudwatch.Alarm.html) - [enum ComparisonOperator](https://docs.aws.amazon.com/cdk/api/v2/docs/aws-cdk-lib.aws_cloudwatch.ComparisonOperator.html) - [enum TreatMissingData](https://docs.aws.amazon.com/cdk/api/v2/docs/aws-cdk-lib.aws_cloudwatch.TreatMissingData.html) - [class Topic (construct)](https://docs.aws.amazon.com/cdk/api/v2/docs/aws-cdk-lib.aws_sns.Topic.html) - [class Subscription (construct)](https://docs.aws.amazon.com/cdk/api/v2/docs/aws-cdk-lib.aws_sns.Subscription.html) - [CDK Construct Library for Amazon Simple Notification Service Subscriptions](https://docs.aws.amazon.com/cdk/api/v2/python/aws_cdk.aws_sns_subscriptions/README.html) - [Metric objects](https://docs.aws.amazon.com/cdk/api/v2/docs/aws-cdk-lib.aws_cloudwatch-readme.html#metric-objects) - [class Metric](https://docs.aws.amazon.com/cdk/api/v2/docs/aws-cdk-lib.aws_cloudwatch.Metric.html) ### Steps to Test <!-- If this work affects a user's experience, provide steps to test these changes in-app. --> 1. Make sure you've authenticated to the correct AWS account using short-term credentials (see README for instructions on how to do this). 2. Run `npm install` to install dependencies. 3. Run `npm run synth:api` to synthesize the Feedback API into a CloudFormation template. 4. Run `npm run deploy:alarm` to deploy the Feedback API, which includes the alarms 5. Verify that the CloudWatch alarms are created as expected using the AWS console --------- Co-authored-by: elizabeth zhang <elizabeth.zhangey@gmail.com>
ezhangy
added a commit
that referenced
this pull request
Aug 26, 2025
Reverts #81 (should have merged with a merge commit, but instead squashed and merged)
ezhangy
added a commit
that referenced
this pull request
Aug 26, 2025
Reverts #82 #82 reverts #81, which I did because I wanted to merge with a merge commit instead of squashing and merging. But afterwards, I couldn't create a new PR to merge dev into main that contained all the code changes we wanted (not sure why). So reverting the revert so we can at least get back the code changes.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Merges the following PRs to main: