ratpik

When a flow is directly triggered by a users action, we have the Cookie/Token of the user to do AuthN/AuthZ.

However when we need to run background jobs or some asynchronous workflow on behalf of the user, we do not have the Cookie/Token of the user. We do have the user id but not the session. How do you solve for AuthZ specifically for your background jobs and dependent microservices?

We are planning to adopt SNS as a mechanism for services to publish events to. We will have 8-10 teams that will adopt this and then scale to 30 teams.

We have been using Postman for the API development of HTTP APIs (debugging, testing, documentation, sharing with other teams).

Is there something similar that helps with the development for APIs that are asynchronous?

Looking to learn from people who have built web/mobile applications on AWS/Azure/GCP at scale (100-200 Million DAUs performing 50:50 write/read operations)

Context: We have a few microservices, each with their own MySQL databases that communicate with each other over HTTP 1.1. A few synchronous operations that constitute a single transaction require writes across two different microservices.

The system can end in states where one of the microservices completes it part of the transaction but the other does not and there is no straightforward way to do rollbacks.

How do you solve this problem?

Trying to figure out how to build something that would return results fast (sub 100 ms) and also support frequent writes.

Data Store - Elasticsearch Scale - 10 million writes/day (500 GB/day), about 100K search queries per day

Trying to figure out

1 - How to control access to data (multi-tenancy where there are ~100K tenants)

2 - Database design - Indexes and Shards and best practices around mixing different types of documents in a single index.

Trying to put together a list of projects written in Node.js that can serve as a reference for best practices in terms of how the code is organized, appropriate usage of core functionality and everything else that can qualify as good architecture.

Traffic = About 0.5 million requests/minute (HTTPS + WSS) and about 80K concurrent websocket connections with intermittent spikes in traffic.

Runs on managed infrastructure on AWS. Almost all the requests are Network (HTTPS) I/O intensive and about a third are Database (MySQL) I/O intensive.

Could be related to software engineering or any other field that is enabled via technology.

Do include a link to the talk if it is available online.

Given a dump of several logs files spanning across different components of your server side stack (like proxy, application server) what is the fastest technique/tool to determine what caused an error?

I have a bunch of services running on AWS EC2 instances (with nginx and docker) managed with Elastic Beanstalk. Whenever someone reports an issue, we need to take a look at the logs for finding errors.

Do you use a third party service (like social auth), an identity provider or use your own? If so what are the pros/cons from your experience?