PyGDA #88 (In-person Edition)

28 September, 18:00 | Pub Lawendowa 8, Gdańsk Main Free entry, registration on meetup.com --- Agenda: 18:00-18:15 - networking + intro 18:15-XXX - Presentation #1: Casper Forge When the cloud starts to burn. Challenges of orchestrating and scaling distributed databases 19:00-19:15 - break XxX-20:00 - Presentation #2: Sebastian Buczyński - Biggest myths on the 20:00 line... - networking --- Casper Forge – When the cloud starts to burn. Challenges of orchestrating and scaling distributed databases.

Description: Transfer of the state base to the cloud and automatic orchestration of thousands of instances sounds like a perfect plan. Dispersed systems are meant to take care of cohesion themselves, and cloud take on failures. Unfortunately, on a scale, much of these assurances become myths.

If you can't sleep at night through noisy nighbors, and with split-brain will not help you any aspirin – this speech is for you. Let's see what happens when cloud mechanisms start fighting the logic of the database, and mass migration jams the network and storage. Without theories and slide requests: pure practice, learning from real failures and seeking a compromise between reliability and productivity.

Bio: Senior Software Engineer at Housecall Pro (Search Platform), ex-Allegro, ex-Revolut. A passion for distributed architecture and efficient backend systems. After hours he trains table tennis and boxing.

Linkedin: https://www.linkedin.com/in/kacperkuzniarski/ --- Sebastian Buczyński – The biggest myths in work with queues Description: Sending messages to queue after the transaction is approved is ok. If the message gets in line, it'll be served. Renews are enough for the message to be finally handled and no man will have to look at it.

Kafka and RabbitMQ differ like day and night, so you have to approach them differently. These are just a few more popular misconceptions about queues, background tasks and messages. Let's see what it's really like looking for...