Web and application performance audit
Monitoring, load testing and optimisation of your applications in production. We measure what actually slows your web applications and websites down, we rank the causes, then we implement the fixes and watch the result over time.
Audit and fixes
Load and stress testing
Monitoring up to 24/7
/ Our Clients
/
What is an application performance audit?
An application performance audit measures how your application actually behaves under load, then identifies what slows it down. The starting point is metrics, not impressions:
- Response times per page and per call, and resource consumption
- The most expensive database queries, and contention points on the server side
- A deliverable that ranks the causes by expected gain and by effort to fix
- Not to be confused with a technical security audit, which looks at vulnerabilities rather than response times
/ Our performance audit approach
Four stages, from the initial measurement to long-term monitoring. Each phase produces a usable deliverable, and the before and after gap is measured on the same indicators as the initial audit.
Initial audit
Measurement of response times, analysis of logs and queries, identification of contention points across the whole chain, from the browser to the database.
Prioritised action plan
Ranking of causes by expected gain and effort to fix, with the optimisations costed and the trade-offs made explicit.
Implementation
Implementation of the fixes, validation testing and measurement of the before and after gap on the indicators from the initial audit.
Continuous monitoring
Monitoring of availability and response times over time, alerts on degradation and regular reviews, up to 24/7 on-call depending on criticality.
Monitoring and optimisation of your applications in production
Beyond the audit, we provide continuous monitoring of your applications and deal with degradations as they arise, with cover ranging from business hours to 24/7 on-call.
Application monitoring and synthetic monitoring
Simulation of user journeys from a browser, at regular intervals, to detect a degradation before your users do and measure real availability.
Log centralisation and analysis
Collection and correlation of application and system logs, to diagnose an incident in minutes rather than hours.
Performance testing and automated QA
Manual and automated testing, web and mobile, built into your release cycles to stop a performance regression reaching production.
Database and slow SQL query optimisation
Analysis of the most expensive queries, review of indexes and execution plans: this is the most common cause of back-end slowdowns.
APM setup and management
APM deployment, instrumentation of your applications and, above all, use of the data: a tool with nobody to interpret it serves no purpose.
/
Load and stress testing
A load test simulates hundreds or thousands of concurrent users on your application, to find out how it behaves under pressure before your customers discover it for you. Four variants answer different questions:
- Load test: do response times hold up at the peak traffic you expect?
- Ramp-up test: at how many users does the application slow down, then give way?
- Stress test: how does the system degrade beyond its limits, and does it recover?
- Soak test: does load sustained over several hours reveal memory leaks?
- Deliverables: curves by user tier, saturation threshold, the bottleneck identified and sizing recommendations
/ Technologies we use
We combine monitoring tools, APM platforms, testing solutions and log stacks, according to your architecture and your constraints.
A stack adapted to your technical constraints and business needs.
/
Our commitments
We guarantee continuous, proactive monitoring of your applications, targeted fixes and an improvement measured on the same indicators as the initial audit. No report without follow-through: we act on the critical points we have identified.
Cover ranges from business hours to 24/7 on-call depending on how critical your services are. More than 25 years of experience and more than 500 projects delivered, with more than 80 employees at Castelis, a hunik group company bringing together 7 specialist firms.
/
Frequently asked questions
It is the measurement of how your application actually behaves under load, followed by the identification of what slows it down: response times per page and per call, resource consumption, the most expensive database queries, server contention points. The result is a plan ranked by expected gain and effort to fix, not a mere observation.
A load test verifies that your application holds up at the peak traffic you expect: it is a validation. A ramp-up test increases the pressure progressively until the application gives way, to determine at how many users it slows down and then breaks: it is an exploration. The first answers « will it cope? », the second « how far will it cope? ».
Application monitoring watches your applications, measures their availability and alerts you when they degrade. Managed services go further: they take on operations, incident handling and fixes, within a contractual framework with service levels. If you have a team able to act on alerts, monitoring is enough. If not, managed services are the right scope.
It depends on the scope: the number of applications, the complexity of the architecture and how deeply the existing stack is instrumented. An audit on a targeted scope produces usable first results quickly, whereas a complete application chain requires more measurement. The initial scoping gives you a firm duration before any commitment.
Yes. Cover comes in three levels: business hours, extended hours and 24/7 on-call. The level is chosen service by service according to actual criticality, rather than applied uniformly: it is the factor that weighs most on the cost of the arrangement.
Yes. We work on websites, business applications and mobile applications, adapting the testing and monitoring tools to each environment.
Let's talk about your application performance
