Wonderful presentation. 3+ years later it is still relevant. - Wonder if periodic broadcasting / pushing of the number of events would be a solution (to catch any users running into the appsync race condition listed around 13:10) - But perhaps that would create a network bottleneck. - Extremely curious to see if any one has come up with a work around solution in 2022. It still blows my mind that a direct data connection from AppSync to Elasticache has not been made available yet. - Also, aren't the lambda scaling limits flexible? Can't you just call them up and ask them to remove the lambda limits? Puzzling through this thought exercise was a huge amount of fun - Thank you!
Really nice presentation! Thanks for the insights in all these services. Since you didn't got ask any questions, I have a few easy ones: - Did you guys considered using Google Firestore service? It provides real time updates to subscribers and it is meant to scale pretty nicely. - In the final architecture, the lambda that triggers after AWS Gateway requests, writes into DB and then publish to an SNS topic. Did you explore using DynamoDB stream subscription to replace SNS topics? It provides a really nice integration with Lambda and out-of-the-box retry functionality. - Is your ECS solution running Fargate or EC2? Thanks and again really nice talk!
