Why Most Code Reviews Miss Performance Issues—and How to Fix Them

Read this: The article discusses key issues that need to be addressed in the code review process during software development.

I almost missed it; this is an extremely important comment.

As the industry matures and code becomes more complex, many organizations run unit, integration, and even end‑to‑end tests, yet few incorporate performance testing—a critical gap that must be addressed.

Code Review – What We Need to Determine

Understand the purpose of code review

Provide polite and actionable feedback

Ensure code meets operational standards

Confirm there are no obvious issues

Maintain open dialogue to keep the process smooth

…

Code Review – Impact on System Performance

As systems grow more complex and teams adopt continuous integration and continuous deployment, the importance of performance considerations in code reviews increases.

Code Review – What We Need to Watch For

Does it introduce an N+1 query problem?

Are there subtle benefits or drawbacks to caching?

Are we getting stuck in a deadlock?

…

Although many organizations already run unit and integration tests in their pipelines, and more are adding end‑to‑end tests, very few include performance testing. This is a gap we need to close as the industry matures.

Remember to consider downstream impact and potential large‑scale issues when pulling data from external services; they may need to query other systems.

These extended problems can appear in unexpected places. Previously, some organizations fetched data asynchronously without issues, but new use cases involving massive parallel calls exposed problems.

When caches miss on the other side, the external system also suffers from repeated data lookups.

Of course, there are many solutions:

If a field cannot be retrieved, discuss with the owner and fetch it manually.

Warm the cache with an initial call.

Align on a base cache strategy.

…

If you don’t monitor performance or conduct load testing, such issues may only surface late, and a single call won’t trigger them.

A Major Challenge: Time

We all experience situations where we need to resolve tickets, perform code reviews, and conduct technical analysis under tight schedules.

That’s why simplifying as many tasks as possible is crucial.

For local analysis, we can use an IntelliJ IDEA plugin to bring telemetry data closer to developers’ workflow.

If we use GitHub, we can leverage two new Digma actions.

Digma Action – instrument

Allows us to add instrumentation using OTEL data and send the captured data to the Digma Analytics Engine.

Digma Action – assert-no-issues

Can be used to verify that no ignored issues are part of a pull request.

https://github.com/digma-ai/digma-actions

Performance Impact: The Gap We Must Bridge

This gap should be addressed during development. Tools like Digma make local code review instrumentation convenient.

While many organizations run unit and integration tests and increasingly add end‑to‑end tests, few include performance testing. As the industry matures, we need to fill this gap.

Author: Simon Verhoeven, senior software engineer and Java consultant, focusing on cloud quality and maintainability.

Related reading:

How to Conduct Effective Code Reviews?

Rich Harris: Svelte 5.0 Reduces Code Volume

Code Review Automation? Amazon Launches New Tool