Consumer Throughput? I don’t care. Kafka Time Lag is my uber-metric for SLO Monitoring.More than a year has passed since I published SLO Monitoring of Event-Driven Architecture which essentially suggested starting with…Jan 7A response icon1Jan 7A response icon1
Identify Job Completion in multi-phase Kafka consumersEven the title raises some questions. If it’s a job that partially has a start and end, why run it over an unbounded stream system like…Jan 11, 2024Jan 11, 2024
The hidden cross AZ cost: how we reduced AWS Data Transfer cost by 80%As expected, AWS services are not for free. And we, as BTB SaaS, are not free either. We definitely try to minimize the infrastructure cost…Oct 26, 2023A response icon2Oct 26, 2023A response icon2
SLO Monitoring of Event-Driven ArchitecturesEvery SaaS journey probably begins with observability, which is the essential capability to gather data from metrics, logs, traces, and…Jul 18, 2023Jul 18, 2023
Published inCodeXHow we (almost :)) achieve End-to-End Exactly-Once processing with FlinkDistributed stateful stream processing is challenging, especially in regard to handling failures and recovery. In stream processing, one…Aug 29, 2022A response icon3Aug 29, 2022A response icon3
Published inCodeXSynchronize Different Consumers of the Same Kafka TopicSynchronization in distributed systems is tough. You probably aim to prevent it as much as you can. But sometimes business requirements…Jul 28, 2022Jul 28, 2022