Daniel Lockyer
I help teams fix high CPU, memory leaks, slow APIs, and rising cloud bills - without rewriting everything.
Performance Retainer Performance and Reliability Retainer
Limited Q2 availabilityWhat this covers
- Production profiling (CPU, RAM, I/O, databases)
- Root-cause analysis of performance and stability issues
- Shipping code and infrastructure changes
- Database and query optimization
- LLM and cloud cost optimization
- Ongoing improvements to prevent regressions
How I work
I focus on the areas with the most impact - not surface-level optimizations. Initially, this means identifying the primary sources of performance, stability or cost issues. From there, I ship fixes directly, and provide proof of their effectiveness.
Most clients see meaningful improvements within the first month: 50-90% reductions in response time, CPU, or costs are typical outcomes.
I work embedded in your team (Slack/team chat), and handle issues end-to-end rather than handing off reports or recommendations.
$12,000/month
Minimum engagement: 3 months.
Scott Baxter
First Internet
We've worked with Daniel since 2018, and he's been an outstanding expert for our team. He dramatically optimised our infrastructure, reducing server load and significantly improving performance and uptime.
His proactive approach means he's often resolved potential issues before they become real problems. Daniel is highly responsive and approachable in his communication.
If you're looking for a skilled, dedicated, and reliable specialist, Daniel is your man!!!
Marek Bron
Editor of Indietraveller
If you're wondering whether you should hire a performance expert, Daniel is your answer. He analyzed my setup and dramatically reduced CPU load (in the 80-90% range). Not only was that faster, but my new hosting solution was cheaper.
Bruce van Zyl
Founder of Tribe Social
Daniel did a fantastic job. Had issues with servers for 4+ weeks, going back and forth with another senior contractor. Daniel had it fully recovered in less than 24 hrs.
Great communication - made it the whole recovery painless experience!
Nicolas Le Roux
Founder of La Langue Francaise
Daniel optimized the performance of my website, which has over 2,000,000 monthly visitors, in a quick and efficient manner. He reduced the CPU load and loading time of the site regardless of the poor efficiency of third-party ads.
I really appreciated his answers to my questions regarding server management and optimization throughout the process.
Definitely recommend!
Highlights
A few more changes and the P50 for my client's API has dropped by 66%
— Daniel Lockyer (@DanielLockyer) December 19, 2025
Not only faster for us, but faster for the users who build on top of the API!
🚀 https://t.co/6gz3zazRka pic.twitter.com/SPHVMUXdqj
I just fixed the underlying issue for the client (app server <-> DB/Redis location) and improved the response time across the app 🎉 https://t.co/9yG4JrLqdp pic.twitter.com/zEQIUO8BCI
— Daniel Lockyer (@DanielLockyer) December 16, 2025
I've just opened a PR to @ClickHouseDB's ClickStack/HyperDX to remove ~80MB of RAM usage in local dev and production
— Daniel Lockyer (@DanielLockyer) December 19, 2025
The problem? Prefixing `npx` where it isn't necessary
Sounds small but you want all RAM going to caching data for query lookups, so it helps! pic.twitter.com/boFlVDmDS0
One of my latest clients came to me with (among many things) a persistent memory usage spike in their Node.js app
— Daniel Lockyer (@DanielLockyer) July 2, 2025
To counteract it, they had bumped their instance size 8x to soak up the excessive RAM usage, and over-provisioned instances to handle stability issues when the OOMs… pic.twitter.com/v7ZncnHrwE
— Daniel Lockyer (@DanielLockyer) November 28, 2025
Another week, another frontend performance bug...@Steen3S and I have found an issue in the search page of @Namecheap, resulting in 80% CPU and >3GB memory usage
— Daniel Lockyer (@DanielLockyer) September 18, 2025
We found the cause to be heavy CSS animations for loading elements which render even after the results are showing… pic.twitter.com/kuQQzgXEo8
Just enabled jemalloc for a client's Rails app on Heroku and their memory usage dropped by ~40% pic.twitter.com/4dldNj0yDN
— Daniel Lockyer (@DanielLockyer) July 9, 2025
Just found some old photos from my consultancy work 5yrs ago.
— Daniel Lockyer (@DanielLockyer) September 13, 2024
This one was a PHP 5.6 to 7.2 upgrade - ~140% CPU to ~10%?
It was one of those times where the graph is so unbelievable that you need to double-check the website is still running (it was) 🙈 pic.twitter.com/vBbiHjBWgw
If you don't have frontend error monitoring, or don't look at it, I'd strongly suggest you do
— Daniel Lockyer (@DanielLockyer) August 6, 2025
It's one of the best ways to improve the UX (less errors/crazy stuff)
In this screenshot, an error caused an extra blocking HTTP request. Fixed and now the frontend is 500ms quicker pic.twitter.com/Ni0OjbURxA
when you make the app so efficient it doesn't even need to autoscale anymore pic.twitter.com/k6QuiT64IX
— Daniel Lockyer (@DanielLockyer) July 17, 2025
Pro tip: ensure you have TLS connection reuse enabled if your app makes a lot of external HTTPS requests
— Daniel Lockyer (@DanielLockyer) July 24, 2025
Just enabled it for a client and their CPU usage has dropped by ~25% w/ lower P75 and P90 response times pic.twitter.com/IaoZkZ42lw
Other Services
Production Review Session
One-off call to get advice on your production challenges. I'll review your specific situation and provide recommendations.