0:00
/
0:00
Transcript

We touched upon performance in the last few editions with:

Still, I’m not a performance expert here, so I invited Jarosław Pałka to our webinar. He is an expert.

He’s the Senior Staff Software Engineer responsible for benchmarking infrastructure in Neo4j.

Jarosław Pałka (@j_palka) / X

Fun fact: He’s also responsible for me becoming a speaker, as he helped me get first on stage in the SegFault conference he was running. He’s one of the most active members of our Polish IT community, so blame him!

Getting back to performance tests, benchmarking and that fun stuff.

We discussed how to reason about performance practically, so how not to get mad and build performance-aware systems. We focused on the strategies and techniques, not on specific tooling. That you can learn in other places. We went beyond a simple micro-benchmarking showcase.

Jarek explained that we have multiple levels of performance tests:

  • benchmarking,

  • latency tests,

  • throughput tests,

  • load tests,

  • stress tests,

  • endurance tests,

  • capacity tests,

  • degradance tests.

He also shared his practical advice on starting with them, not getting mad, and not being too afraid even to try that.

We also discussed marketing and how to apply lessons from Neo4J performance testing to the regular systems!

See also more in the materials Jarek recommends:


Webinars: