Speedshop

Performance monitoring, roadmapping, and support.
100% managed by Speedshop.

Do you ever stare at your APM dashboard and wonder: what the hell did I install this for? You don't understand 90% of the numbers, and if you were asked to find important metrics such as time spent waiting on GC, time spent waiting on I/O, the 95th percentile latency of your external services, or others, you wouldn't have any idea where to even start.

Slow site, but you don't know where to start?

You know you want to improve the performance of your Ruby web application. Perhaps customers are complaining, or churning out because they feel like they're "too big" for your platform these days. You've got a solid product engineering team, but maybe you don't feel like your operational chops are quite there yet. In any case, you'd rather that your current engineers focus on delivering features, and only spend the minimal amount of time on perf work required to put out the fires and move on.

Or, maybe you feel like the lights just aren't staying on. Every time a new big client gets brought on, or an influx of traffic hits the app, everything starts to "brown out" like the local power plant just fizzled in and out. You might have a bandaid for today, but why does this keep happening?

Have you ever felt like every time you actually tried to fix a performance problem, it was like quicksand for your engineers? Days become weeks, weeks may even become months, and yet there's no concrete progress, and customers are still complaining?

Replicating a proven playbook at your business

From 2022 to 2024, Speedshop was able to reduce page load times at Gusto by over 40% across the entire application. That's a massive improvement for a multi-billion dollar business with hundreds of Rails engineers. For businesses of any size, we can take the relevant parts of that proven process and bring it to your organization.

41 businesses have already joined this program since its creation in 2024, and seen similar results. We'll work with you to identify what key metrics are important to your business, and then make that our goal to improve over the 6 month contract period.

What we've done in the last 365 days

These 741 individual pieces of work are logged and updated daily from our internal notes.

334 Code

PRs that make code faster, PR reviews, env variable tuning, custom library creation.

Pull requests opened
A pull request against your app or infrastructure that makes things faster, more scalable, or more observable.
Pull requests updated
Revising open PRs, such as responding to review, rebasing, or making CI green.
Pull requests reviewed
A performance-focused review of a pull request written by your team.
Production changes
A change applied directly to production configuration or infrastructure, such as server settings or autoscaling.
Tooling
Building or improving a Speedshop library or tool used in your app, such as benchmark harnesses and instrumentation gems.
151 Office hours

A recurring call with your engineers to review metrics, plan work, and answer questions.

Office hours calls
A recurring scheduled call with your engineers to review metrics, plan work, and answer questions.
82 Diagnosis & research

Root-cause analysis of perf issues in prod, creating benchmarks.

Investigations
Diagnosing a production incident, regression, or performance problem down to its root cause.
Research
Benchmarks, experiments, and tool evaluations that answer a performance question before any code changes.
76 Observability

New and improved dashboards, metrics, and SLOs that measure what matters and catch regressions.

Dashboards created
A new dashboard tracking the metrics that matter for your app's speed, scalability, and stability.
Observability updates
Improving an existing dashboard, metric, monitor, or SLO so it measures the right thing.
SLOs created
A new service level objective with a target and alerting, so regressions are caught automatically.
62 Planning & reporting

Triaging your performance related tickets, doing audits, and directing work.

Issues created
A new ticket in your tracker that describes a performance problem and how to fix it.
Issue comments
Analysis, measurements, or direction added to an existing ticket.
Issues completed
We successfully fixed something and documented its impact on prod.
Backlog triage
Organizing, prioritizing, and pruning the performance backlog in your tracker.
Audit reports
A written performance audit of your application with prioritized recommendations.
Monthly reports
A monthly summary of key metrics, progress, and next steps.
36 Chat & calls

Async answers in team chat, unscheduled calls (not all Slack messages count here, just the big/important ones).

Chat conversations
An async conversation in your team chat: answering a performance question, advising on a change, or following up.
Ad hoc calls
An unscheduled call, such as pairing on a problem or talking through a design decision.

An optimized app, delivered continuously

There are three critical components to any successful performance effort:

  1. Extensive and correct instrumentation. You need to have the tooling and observability installed correctly and measuring the numbers which need to be measured. It goes far beyond just adding an APM to your Gemfile and calling it a day.
  2. Sufficient performance skills. Speedshop has taught hundreds of Rails developers how to optimize Rails apps, through workshops, books, and on-site training. Through this ongoing service, we can provide targeted training where required.
  3. Clear technical direction. Metrics are good, but you need to turn them into concrete work: stories, tickets, or whatever your existing system is, we'll turn performance metrics into work that needs to be done, and provide context on priority and difficulty of implentation.

This service provides all three. And, it's all customized to your specific situation. Scalability got you down but page latency looks ok? Sounds good, we'll focus on that. Sidekiq a dumpster fire but the web process looking fine? Let's fix your background jobs!

What does the service entail?

This is a monthly retainer that provides a performance product owner for your organization for the purpose of optimizing performance and scalability.

You'll give Speedshop access to your performance metric provider, such as New Relic, Skylight or Scout, and any other server metrics. We'll also need read (not write) access to your source code. We'll discuss your particular performance needs - need a tuneup for the frontend or backend (or both)? Do you have memory issues? Are a few particular workers or controller actions cramping your style?

Once that's done, here's what we provide:

  1. Performance monitoring and roadmap management, 100% managed by Speedshop. Monitoring performance issues, creating tickets/roadmaps or otherwise "scoping work" based on these, will no longer be your problem. You should feel like the assessment of perf related concerns and identification of "what needs fixing and how to do it" is no longer on your plate. You'll also know that any perf work you do will actually be successful and have customer impact - no more barking up the wrong tree for weeks on end and ending up with half-solutions.
  2. Setup of automated alerts and dashboards.
  3. Monthly reporting with comments from Nate.
  4. Biweekly office hours for synchronous discussion with Nate.
  5. Unlimited pull request review for perf-related work.
  6. Unlimited async access for perf-related issues via Slack/team chat.
  7. Your entire team will get access to all of my written products and books plus the Rails Performance Workshop, a $600 value per person.

Let's get started.

This ongoing retainer service is available for $3,000 USD, monthly. Contracts are a six-month term. Many of our customers end up renewing.

Availability

December January February
❌ ❌ ❌

Legend:

  • ✅ Available - 3+ spots open
  • ⚠️ Limited - 1-2 spots remaining
  • ❌ Full - No spots available

Start the retainer now.

Let Speedshop own your Rails application's performance. Book a call to discuss your app, confirm availability, and get started with ongoing monitoring, roadmapping, and support.

Who's this?

My name is Nate Berkopec. I'm a software developer and author of the Complete Guide to Rails Performance. I've previously worked for people like Gusto, Intercom.io, and more YCombinator startups than I can count. I was on Shark Tank once.

Nate Berkopec