Performance audit: find your system's limit before your users do
In 1-2 weeks we locate bottlenecks, measure the real limits and hand over a prioritized optimization plan. No guesses — only p95/p99 metrics, profiling and scenarios that mirror your traffic.
Timeline
1-2 weeks
Pricing
Fixed price
Output
Report with metrics, bottleneck map and action plan
What you get
Four concrete artifacts that stay with your team and keep working long after we leave.
PDF report with p95/p99 latency, throughput and error rate for key endpoints — measured under load close to production traffic.
Bottleneck map showing where you hit the database, where CPU or memory tops out, where the network or code is the limit — down to specific queries and methods.
k6 (or Locust) scenarios and Grafana dashboards we built — yours to keep, so you can run regression checks before every release.
Optimization roadmap: what to fix first, which changes give +30% throughput, and what needs an architectural decision rather than a quick patch.
When to order
If you recognize your situation in any of these — it's time to measure, not guess.
Before a marketing push, TV campaign or seasonal peak — so the predictable spike doesn't take you down.
After 3-5× user-base growth, when your old assumptions about load no longer hold.
Before expanding to a new region or market — you need to know the infrastructure will keep up.
When your APM shows p95 degradation but it's unclear what's slowing things down — you need a systemic profile, not point hypotheses.
How we run the audit
Four phases in 1-2 weeks — from getting to know the system to a report with a prioritized plan.
01
Context and goals
We study the architecture, business scenarios, current metrics and SLAs. Together we agree which peaks and scenarios to simulate.
02
Profiling
We capture the baseline: latency, throughput, resource usage. We dig into APM, logs, database and cache metrics.
03
Load scenarios
We write k6 scenarios that mirror real traffic patterns. Run step-load, soak and spike tests. Document failure points and limits.
04
Report and prioritization
We aggregate findings, score the impact of each optimization, present results to your team and hand over all artifacts.
Pricing and scope
Exact figure — after a short discovery call. These ranges are for the KZ market 2026, depending on scope.
Light
800K–1.5M ₸
1 service, key endpoints, baseline + 2-3 load scenarios.
Standard
1.5–2.5M ₸
Multiple services, DB profiling, spike/soak/step-load scenarios, recommendations.
Deep
2.5M+ ₸
Distributed system, multi-service profiling, migration scenarios, on-call during peaks.
Related products
Often ordered together
Discovery: from idea to spec, budget and timeline in 2–4 weeks
2-4 weeks
Learn More
Software architecture review: an outside view while change is still cheap
1-2 weeks
Learn More
Infrastructure audit: what you pay the cloud for and where the risks are
1-2 weeks
Learn More
Questions & Answers
Frequently Asked Questions
Measure the system before your users do it for you
Tell us about the product and your load scenarios — within a day we'll come back with an audit plan and timeline.