October 29, 2025
Keep Calm and Use Postgres
Kafka is Fast – I'll use Postgres
Commenters roast the buzzword craze and cheer the simple choice
TLDR: A developer argues most teams should stick with Postgres instead of deploying Kafka for small workloads, sparking a chorus of “keep it simple.” The top reactions praise practicality, roast “resume-driven” hype, and add caution: Postgres is great—just don’t use it for absolutely everything.
A spicy blog post declared: “Kafka is fast — I’ll use Postgres,” and the internet lit up like a data center. Translation for non-geeks: Kafka is a heavy-duty data conveyor belt; Postgres is your trusty Swiss‑army database. The author says most teams don’t need a rocket ship for grocery runs, and the crowd largely yelled “amen.” The loudest cheer? The “keep it simple” camp. One commenter confessed, “I feel so seen,” calling out hype tech left lying around “for someone’s résumé.” Another dropped the meme of the day: ask if the person who pushed the fancy system still works there—spoiler, they usually left for a shiny new job.
There’s drama, too. People are dunking on “resume-driven design” while rallying around the “Small Data” vibe and a full‑blown Postgres renaissance: why add five tools when one can handle 80% of the job? Fans shared links like pgflow.dev with a mic‑drop: “Postgres is enough.”
But the caution club showed up. Veterans warned that using Postgres for literally everything can cause headaches—like locks (when the database makes sure things don’t clash) and slowdowns if you pile on too much. Still, the mood is clear: for modest workloads, don’t overbuild; pick the boring, reliable stuff and go ship something.
In short: hype stacks got roasted, practicality got applause, and Postgres took a victory lap.
Key Points
- •The article advocates defaulting to PostgreSQL for most workloads, emphasizing simplicity over hype-driven architectures.
- •Two trends support this view: typical data sizes are modest while hardware has grown powerful, and the Postgres ecosystem is rapidly expanding.
- •Postgres features and extensions can cover many use cases of specialized systems like Elasticsearch, MongoDB, Redis, AI vector databases, and Snowflake.
- •Kafka is acknowledged as mature and scalable but is often unnecessary for small workloads; a 500 KB/s workload is cited as an example.
- •The article proposes benchmarking Postgres for pub/sub messaging and queueing, and discussing when Postgres is an appropriate fit.