1 March 2022 – 31 May 2023
Migrated Orleans from 3.1.7 to the latest version.
Services (silos) can run jobs, and any job with given key can be run at most once at any moment of time. Orleans promises that the data is not corrupted (modified concurrently by different instances of services). Assessed risks, ensured we can rollback, created migration plan. Within four weeks learned new technologies [originally, I worked with Java stack]. Solved problems with services communication (clustering – gossip communication in the same datacenter; multi-clustering – gossip communication of services in different datacenters) – at the beginning of migration services could not discover each other. Fixed compatibility issues (new version of the framework changed the algorithm for id generation), added tests to ensure records ids are the same with newer versions of the library. Finally, deployed to production and tested the correctness (services discover each other, data is not corrupted). More details and problems I had to solve https://github.com/yan-khonski-it/orleans-migration-experience/blob/main/README.md
Resolved 60+ incidents during on call weeks, deployed a hotfix, created tasks to address false alarms. All this during my two months of work.
Tech stack:
- Azure
- Azure Traffic Manager
- CosmosDB
- .NET
- Orleans
1 January 2022 – 28 February 2023
Internally moved to Wrike after Citrix restructuring at the end of 2021.
Reduced the number of queries on the database by 10 %, while our DAU grew by 2 % during Q1 2022. Analyzed most frequently executed queries, added caches (distributed, local, layered), found and got removed unnecessary functionality with extra load, and changed the logic to reduce the number of calls.
Resolved production incidents. An example, we had 3000 request per minute for one public API end-point from one customer. As it turned out, the end-point would invoke a database query, which timeout the client request, so the client would retry. Added rate limiter for our public API end-points based on customers plans and user settings.
Tech stack:
- Java 11, Java 17, Spring, MyBatis
- PostgreSQL
- Redis, internal caches (Guava)
- JMS (RabbitMQ)
- Kafka
https://docs.citrix.com/en-us/citrix-microapps.html
1 September 2020 – 31 December 2021
Designed, wrote architecture concept, documented (diagrams.net, confluence) "multi-tenant migration of job processing nodes". Organized meetings with related teams, split the problem into manageable tasks and assigned them to the teams (and people).
Problem: The monolith code base was single tenant application. We had thousands of cloud instances (each tenant required multiple instances for high availability). The motivation was to reduce the running cost.
We decided to migrate the job processing functinoality to make it multi-tenant and scalable. The next step would be to migrate the rest functionality. Result: during my implementation phase, Citrix laid off its employees of the whole business unit...
Previous tasks.
Fixed: performance issues (Hibernate N + 1 problem, re-wrote the code to use more efficient queries), distributed deadlock (database update and cache update waiting for each other!).
Fixed build pipeline issues, flaky tests.
Increased test coverage of business logic (and inspired other do so) from 60 to 90 % for some modules. Added contract (Pact) tests. Added configuration tests which ensures that configuration and properties files are valid and application can work.
Convinced other developers to fix circular dependencies of software components.
Implemented centralized configuration (before different parts of configuration were stored in many places: environment variables, database, command line arguments).
Tech stack:
- Java 11, Spring, Hibernate, MyBatis
- MS SQL
- Redis, internal caches (Caffeine)
- Azure cloud (service bus, event hub, compute instances, etc)
- Docker, Docker compose
- Pact (contract tests broker)
- Mockito
- Testcontainers (for integration tests that needed running MS SQL container)
minor experience:
- kubernetes, helm
23 January 2019 – 31 August 2020
Owned development of our platform consisting of different services written in different tech stacks (Java, .NET, Python, PHP).
Mentored people and conducted knowledge transfers, made lectures for the company (topics: Kafka, services integration, etc).
Fixed stability issues, memory leaks, made performance optimization (example N + 1 problem in database), refactored poor-quality code into easy to read, tested and documented code.
Fixed problems which caused data loss and data inconsistency between services. Examples: improper transactions handling, backward compatibility issues with messaging.
Improved REST API for some services (made it more client oriented instead of database entities oriented). Wrote full regression REST API tests for some components (jMeter).
Tech stack:
- Java 11, Spring Boot (with Apache Tomcaqt embedded)
- MySQL
- MongoDB
- Kafka
- Spring Cloud (Service discovery Eureka, Configuration Service, Gateway).
- Groovy, Spock
The company was bought by Broadcom
https://www.broadcom.com/products/software/continuous-testing/service-virtualization
2 May 2017 - 30 September 2018
Fixed a number of defects and bugs (some of them were critical) which improved customers’ satisfaction.
Led a group of smart interns (10 students) who built a system (wellness and benefits, event reservation for employees). Helped them with Architecture, deployments and hosting.
Tech stack:
- Java 8, Spring
- CSS, HTLM, Javascript
- MySQL
Minor stack:
- Python3, Django, MySQL, Apache Web Server, Windowns Server 2012
https://www.ita-dev.com/?lang=en
20 January 2015 - 18 November 2016
Prototyped functionality and features, so our customer could better understand how the software will work in the future, provided technical advise to the customer.
Automated manual process of filling data (NDA for this project, sorry):
- Wrote a web scroller with a parser that fetches information from the internet.
- Wrote a processor that converts the information into our business records.
- Optimized storing those business records into the database.
- Added a UI application to fetch and display those records to the user.
Fun project (I wanted to visualize our code base and show it to the team): wrote a tool that builds an oriented graph representing classes (Spring Beans) dependencies of the source code directory.
Tech stack:
- Java 8, Swing, Hibernate, Spring, ZK
- CSS, HTLM, Javascript
- Yandex Maps, Google Maps API
- MySQL
- Apache Tomcat
- Windows Server 2012
October 2012 - July 2013
Worked on internal and clients projects (Real estate domain, REIT).
Tech stack:
- Java, Hibernate, Spring
- CSS, HTLM, Javascript
- Yandex Maps, Google Maps API
- PostgreSQL
- Apache Tomcat
1 September 2010 - 30 June 2015
Math and Computer Science degree 5 years (between Bachelor and Master)
Math, more math, and even more math. Played chess tournaments. Had a good time!