Performance and architecture diagnostic

In two weeks you’ll know where your backend loses performance and what to do first

A performance and architecture diagnostic of a system in production, with a fixed scope and a reference price of USD 2,900 for one system and one business question; the exact price is set in the proposal, according to the system and the observability available. You get a report prioritized by impact and effort, improvements your team can apply right away, and a roadmap it can execute without us.

Request the free 30-minute discovery call

On that call we confirm whether the diagnostic fits your case and what observability you have. If it doesn’t fit, we tell you right there.

What it is for

  • Your cloud bill grows faster than the business and nobody knows how much is real load and how much is inefficiency in queries, algorithms, or caching.
  • Your p99 latency has a life of its own: the average looks fine, but the long tail produces timeouts and retries.
  • Your team fights fires without a map: there are several theories about the bottleneck and no measurement to rank them.

The diagnostic replaces those theories with measurements, and the measurements with a plan in execution order.

What you get

  1. A technical report prioritized by impact and effort. Each finding with its evidence (measurements, traces, relevant fragments), its estimated impact on latency or cost, and its implementation difficulty. Only what applies to your system.
  2. Immediate improvements: the ones your team can apply in less than a week, with concrete instructions.
  3. A 30-, 60-, and 90-day roadmap for the structural changes, with dependencies and risks.
  4. A 90-minute delivery session with your technical team, recorded if you authorize it.
  5. Two weeks of email follow-up for implementation questions.

If the analysis concludes that intervening is not worthwhile, the report says so with its evidence: that is a useful result too.

What we need from you

The diagnostic is one week of our engineering spread over two calendar weeks. What sets the pace is access: the clock starts when we have the access and the business question, not when we sign. If an instrumentation week is needed, it comes first, and the diagnostic clock starts when it ends.

  • A technical contact with five or six hours during the diagnostic: kickoff, validation, delivery session, and preparing access and information.
  • Read-only access to the observability you already have: metrics and dashboards (Prometheus and Grafana, Datadog, New Relic, or similar), traces if you have them, and error logs.
  • Read-only access to your cloud and its billing. On AWS, a read-only IAM role for CloudWatch, Cost Explorer, and resource descriptions; on Google Cloud, Monitoring and Billing viewer; on Azure, Monitor and Cost Management reader; on Kubernetes, read access to metrics and manifests.
  • The architecture: diagrams or a description of components and deployments and, if you authorize it, the code under NDA.
  • The business question the report must answer: what decision you want to make with it.

The first step is always the same: comparing the architecture you describe with the one the metrics show.

How it works

  • Discovery call, 30 minutes, free: we confirm the fit, the available observability, and the question. We sign an NDA if you need one.
  • Access: we send you the list of what we need; once it is ready and we have the business question, the clock starts.
  • First week, analysis: a 60- to 90-minute kickoff; architecture compared with metrics; latency percentiles per endpoint, cache behavior, queries and execution plans, contention and concurrency, sizing and cloud billing.
  • Second week, validation and delivery: a short call to check preliminary findings and discard false positives; final report and delivery session. Then two weeks of email follow-up.
  • If at the kickoff we see that instrumentation is missing to answer the question, we tell you that day and you choose: a separate instrumentation week at a fixed price, or a custom diagnostic. The instrumentation week comes before the diagnostic: its two weeks start when it ends.

Price and terms

The amounts are references: the exact price is set in the proposal, according to the system and the observability available.

  • USD 2,900 for one system and one business question, with an electronic invoice: in Colombian pesos for clients in Colombia, in US dollars or euros for clients in other countries.
  • USD 2,200 with an authorized reference: if you agree to let us publish an experience entry we agree with you, with your company’s name or without it, the reference price drops to USD 2,200. The entry describes real results and does not oblige you to a favorable review; you choose in the proposal.
  • Optional instrumentation week: USD 1,200. Only if the kickoff shows that the minimum needed to measure is missing; the decision is yours.
  • Payment: half when booking and half on delivery of the report.
  • Start date: agreed on the discovery call according to available capacity.
  • If after the kickoff we see that the diagnostic does not apply to your system, we cancel it and refund your deposit.
  • If you later want us to implement: the price of the diagnostic is deducted from any project contracted within the following 60 days, up to 25% of the project’s value.
  • Just one specific question? A 4-hour session for USD 600 as a reference, without a report.

General terms are in How we work.

It is not for you if…

  • your product is not in production yet (if you have an idea to build, see the agile MVP);
  • you are looking for extra hands for your team without a defined scope (we don’t lend people to your team);
  • you want a report that validates decisions already made.

Why us

The prior experience of our team members in systems where milliseconds matter is in Experience, and the open-source cache protocol we explain on the blog has been in production since 2020.

Frequently asked questions

What if we have no observability? We find out on the discovery call. If the kickoff shows that the minimum needed to measure is missing, you choose the instrumentation week or a custom diagnostic. The diagnostic’s two weeks start when the instrumentation is done and we have access.

Do you need access to our code? Ideally yes, but it is not essential: with metrics, architecture, and working sessions with your team we get far. Everything under NDA and with read-only access.

Our stack is not C++, Go, or Rust. Does it apply? Yes, when the scope and the available expertise allow it: the diagnostic is about architecture, data, and performance, and it can evaluate systems on the JVM, Node, Python, PHP, or Elixir. If a recommendation required expertise we don’t have, we would tell you.

Is this bait to sell me a big project? The report is written so your team can execute it alone. If you decide to hire us to implement, the diagnostic is deducted from the project; the decision stays on your side.

How does the discount work if we later hire you for a project? The price of the diagnostic is deducted from any project contracted within the following 60 days, up to 25% of the project’s value. For example, on a USD 8,000 project USD 2,000 is deducted; to deduct the full USD 2,900, the project must be at least USD 11,600. If you paid the price with an authorized reference, what you paid is deducted.

How do you invoice? Electronic invoice: in Colombian pesos for clients in Colombia, with the amount fixed in the proposal; in US dollars or euros for clients in other countries, paid by international bank transfer.

Request the discovery call or write to us at contact@simylresearch.com.