Alice Kaerast
Senior Automation Engineer · Infrastructure Tribe
Worked at Sky Betting & Gaming 2012–2019
Alice played a key role in building our Hadoop-based data platform as a DevOps Engineer and then Architect, and then applied this engineering experience in the infrastructure tribe on Apache Kafka and other shared platforms.
Articles by Alice
Kafka on NFS
There is a general recommendation against running Apache Kafka on NFS storage, but nobody really gives a good explanation as to why. In this post we look at some broker crashes we have seen happening on Kafka clusters which use NFS storage and why they were happening.
5 minute read
JMX Metrics in Kafka Connect
The use of JMX metrics in Java applications is often poorly documented and is a feature that people are often unaware of. In this post we explore how to use the JMX metrics provided by Kafka Connect.
11 minute read
Big Data Technology Warsaw Summit 2018
A report from Big Data Technology Warsaw Summit 2018
14 minute read
What we learned at nonbinary.tech
This weekend the first ever nonbinary in tech event took place in London, and I had the privilege of attending.
5 minute read
How we broke Hadoop by optimising services
We’ve been optimising the allocation of services in our Hadoop cluster recently. It turns out a quiet Hadoop gateway server is a bad one.
3 minute read
Towards a realtime streaming architecture
Outline of the streaming architecture we are standardising around in the data tribe at Sky Betting & Gaming
7 minute read
Our Top 10 Big Data News Sources
Keeping on top of an area of technology that is as rapidly moving as the big data ecosystem is hard. Our data tribe share some of their resources for keeping up to date.
5 minute read
Measuring Impala performance using Apache JMeter
Our web performance teams regularly use JMeter to load test our websites to identify performance of the various components involved, but it turns out you can actually use it to directly test the performance of a Hadoop datawarehouse.
2 minute read
Google Phone Numbers in Spark
Our CRM team rely on having clean phone numbers to push SMS messages to customers, various people have tried creating some logic for this validation but surely this is a solved problem.
15 minute read