Amazon S3

Our base infrastructure is composed of Debian based servers running in Amazon EC2 , asset storage with Amazon S3 , and Amazon RDS for Aurora and Redis under Amazon ElastiCache for data storage.

We are starting to work in automated provisioning and management with Terraform.

Sometimes #ad-blocking addons can cause a real headache when working with JavaScript apps. Onboarding assistants (Appcues + elevio ), chat (Intercom) and product usage insight (Hotjar) have all landed on their blacklists. I guess there is a perfectly good reason for this that I just don't know.

In order to fix this, we had to set up our own content delivery service. We chose Amazon CloudFront and Amazon S3 to do the job because it has a good synergy with Heroku PaaS we are already using.

So if you look through my decisions you will see I recently wrote a decision about moving from Netlify to Buddy and Amazon S3.

I want to write another decision saying that I tried this out and actually moved back to Netlify. Buddy was great until they deleted my account and all my pipelines I setup without warning me because I didn't login for a month.

Netlify is amazing and way easier to setup, support is great and they have so many amazing options... I did learn things about Amazon S3 by moving over to there but I'm sticking with Netlify for the long run now.

We are in the process of building a modern content platform to deliver our content through various channels. We decided to go with Microservices architecture as we wanted scale. Microservice architecture style is an approach to developing an application as a suite of small independently deployable services built around specific business capabilities. You can gain modularity, extensive parallelism and cost-effective scaling by deploying services across many distributed servers. Microservices modularity facilitates independent updates/deployments, and helps to avoid single point of failure, which can help prevent large-scale outages. We also decided to use Event Driven Architecture pattern which is a popular distributed asynchronous architecture pattern used to produce highly scalable applications. The event-driven architecture is made up of highly decoupled, single-purpose event processing components that asynchronously receive and process events.

To build our #Backend capabilities we decided to use the following: 1. #Microservices - Java with Spring Boot , Node.js with ExpressJS and Python with Flask 2. #Eventsourcingframework - Amazon Kinesis , Amazon Kinesis Firehose , Amazon SNS , Amazon SQS, AWS Lambda 3. #Data - Amazon RDS , Amazon DynamoDB , Amazon S3 , MongoDB Atlas

To build #Webapps we decided to use Angular 2 with RxJS

#Devops - GitHub , Travis CI , Terraform , Docker , Serverless

To provide employees with the critical need of interactive querying, we’ve worked with Presto, an open-source distributed SQL query engine, over the years. Operating Presto at Pinterest’s scale has involved resolving quite a few challenges like, supporting deeply nested and huge thrift schemas, slow/ bad worker detection and remediation, auto-scaling cluster, graceful cluster shutdown and impersonation support for ldap authenticator.

Our infrastructure is built on top of Amazon EC2 and we leverage Amazon S3 for storing our data. This separates compute and storage layers, and allows multiple compute clusters to share the S3 data.

We have hundreds of petabytes of data and tens of thousands of Apache Hive tables. Our Presto clusters are comprised of a fleet of 450 r4.8xl EC2 instances. Presto clusters together have over 100 TBs of memory and 14K vcpu cores. Within Pinterest, we have close to more than 1,000 monthly active users (out of total 1,600+ Pinterest employees) using Presto, who run about 400K queries on these clusters per month.

Each query submitted to Presto cluster is logged to a Kafka topic via Singer. Singer is a logging agent built at Pinterest and we talked about it in a previous post. Each query is logged when it is submitted and when it finishes. When a Presto cluster crashes, we will have query submitted events without corresponding query finished events. These events enable us to capture the effect of cluster crashes over time.

Each Presto cluster at Pinterest has workers on a mix of dedicated AWS EC2 instances and Kubernetes pods. Kubernetes platform provides us with the capability to add and remove workers from a Presto cluster very quickly. The best-case latency on bringing up a new worker on Kubernetes is less than a minute. However, when the Kubernetes cluster itself is out of resources and needs to scale up, it can take up to ten minutes. Some other advantages of deploying on Kubernetes platform is that our Presto deployment becomes agnostic of cloud vendor, instance types, OS, etc.

#BigData #AWS #DataScience #DataEngineering

Hi, I'm building a machine learning pipelines to store image bytes and image vectors in the backend.

So, when users query for the random access image data (key), we return the image bytes and perform machine learning model operations on it.

I'm currently considering going with Amazon S3 (in the future, maybe add Redis caching layer) as the backend system to store the information (s3 buckets with sharded prefixes).

As the latency of S3 is 100-200ms (get/put) and it has a high throughput of 3500 puts/sec and 5500 gets/sec for a given bucker/prefix. In the future I need to reduce the latency, I can add Redis cache.

Also, s3 costs are way fewer than HBase (on Amazon EC2 instances with 3x replication factor)

I have not personally used HBase before, so can someone help me if I'm making the right choice here? I'm not aware of Hbase latencies and I have learned that the MOB feature on Hbase has to be turned on if we have store image bytes on of the column families as the avg image bytes are 240Kb.