PGConf.Russia 2025

Many developers often face performance issues in the systems they develop. One common solution for optimizing slow business processes is parallelization. But what do you do if the bottleneck is the data insertion into the database, which needs to maintain atomicity?

In this talk, I’ll explain how to speed up data insertion by parallelizing the process in Spring, while ensuring the atomicity of the entire operation. We'll cover batch updates in Spring and PostgreSQL, discuss why updates are heavy operations, and explore ways to speed up the process in the current tech stack. Additionally, I will present other approaches to maintaining atomicity and demonstrate their differences in benchmarks.

This will be useful for practicing engineers.

In PostgreSQL performance testing, benchmarks measure query execution time (latency). To get more reliable results, queries are executed repeatedly, generating a dataset of latency values. Performance is often assessed using standard metrics like the median or mean, but we propose a more advanced approach.

In practice, latency distributions are often multimodal, consisting of multiple underlying distributions with distinct characteristics. In such cases, traditional statistical methods are insufficient, requiring a more detailed analysis of the dataset’s structure.

Our work presents a tool that automatically performs statistical analysis of benchmark results, accounting for dataset-specific features. It detects multimodality, identifies the number and boundaries of dominant modes, and determines key distribution parameters—providing deeper insights into PostgreSQL performance variations.

Poolers play a crucial role in PostgreSQL database operations. In this talk, we’ll discuss the challenges we faced when using and implementing existing poolers and the benefits we gained by developing our own pooler.

This talk addresses a common issue and touches a bit on application architecture. Temporary tables in database applications are typically used for several purposes. 

Firstly, they are used to store intermediate results when implementing complex data processing algorithms. Secondly, in the case of stored procedures, application servers often place large datasets into temporary tables when they are too large or inconvenient to pass as parameters to stored procedures. 

The handling of temporary tables in different DBMSs is implemented in various ways, which often complicates migration from these systems to Postgres. 

The drawbacks of temporary tables are well-known, which leads to a reasonable desire to replace them, where possible, with other methods that can achieve the same goal. This talk focuses on alternative mechanisms provided by Postgres to solve this problem.