Write to us
By clicking on the button, you consent to the processing of personal data and agree to the privacy policy.
Software Testing
We have a large and strong testing team; our testers’ experience ranges from 3 months to 10 years. For every project, we usually build up teams of 1–5 specialists.
Testing confirms that your software works entirely as expected, without bugs. The product is compliant with the Terms of Reference and meets the customer’s expectations and objectives.
– Good programmers must be coding without bugs right away!
Alas, this is not true. Even the best programmers commit mistakes or cannot foresee everything. Mistakes occur due to human factors. The larger the project, the more people are in a team, therefore, the more space for inconsistencies.
The main causes of bugs:
– So, let the programmers test!
Yes, they can. Surely, they test their code to run as stated in the documents, in a laboratory setting. However, it is advisable to entrust further testing in real conditions to professionals. First, a tester’s services are less expensive than the programmer’s, while testing takes a long time. Second, a programmer is confident in his or her product, and he or she will not “dig deep,”. An QA specialists, on the other hand, does not believe in the programmer’s infallibility and will definitely find bugs.
– Good programmers must be coding without bugs right away!
Alas, this is not true. Even the best programmers commit mistakes or cannot foresee everything. Mistakes occur due to human factors. The larger the project, the more people are in a team, therefore, the more space for inconsistencies.
The main causes of bugs:
- Differences in programming language versions;
- Task misunderstanding;
- Special conditions of use;
- Integration with other modules or external systems.
– So, let the programmers test!
Yes, they can. Surely, they test their code to run as stated in the documents, in a laboratory setting. However, it is advisable to entrust further testing in real conditions to professionals. First, a tester’s services are less expensive than the programmer’s, while testing takes a long time. Second, a programmer is confident in his or her product, and he or she will not “dig deep,”. An QA specialists, on the other hand, does not believe in the programmer’s infallibility and will definitely find bugs.
The types of testing we carry out:
Manual testing;
Configuration testing;
Usability testing;
User interface testing (UI testing);
Mobile apps testing;
Component / unit testing;
Integration testing, both internal and external;
System / end-to-end testing;
Users testing.
Automated testing, autotesting;
Semiautomated testing;
Functional testing;
Performance testing;
Load testing;
Stress testing;
Stability / endurance / soak testing;
How is the testing process organized?
We develop testing scenarios
Based on test cases, we develop testing scenarios by feature.
To check the developed software compliance with the initial requirements, we develop positive and negative test cases.
Based on functional scenarios, we develop scenarios for smoke and regression tests.
We conduct smoke and regression tests after every internal release. All inconsistencies are recorded in the bug tracker; after fixing bugs, we carry out Confirmation Testing and record it in the regression test scenarios.
After internal functional and integration testing, we perform a demonstration of work and acceptance testing based on business scenarios provided by the customer (or developed by us and agreed upon with the customer).
Functional testing scheme:
— esting the individual modules (features) of the system;
— Testing the interaction between modules (integration internal testing);
— Testing the interaction with external systems
(external integration testing)
(external integration testing)
— esting the whole system (including a smoke test minimum)
Testing is carried out simultaneously with the development, with a small time shift; it is started upon the system modules’ readiness.
Simultaneously with functional testing, we perform configuration testing for the adequacy of the system functioning in the declared environments (browsers, OS, resolutions).
We review the project documents
We develop smoke test scenarios
All tests are always executed in the same way; that is, the “human factor” is excluded. A tester will not miss a test by negligence or make mistakes in the results.
An automated script shall check the instructions and documents; this allows for a significant reduction of the execution time.
Once automatic scripts are written, their support and the analysis of results usually take less time than performing the same amount of testing manually.
During test execution, a test engineer can do other useful things, or tests can be executed outside of business hours (this method is preferable due to a lower load on local networks at night).
Repeatability
Fast execution
Lower support costs
Execution without intervention
Advantages of Automated Testing
We can configure automated testing, though it will not suit everyone’s needs. Automated testing is suitable for companies that develop their own products. It makes no sense to introduce automated testing into the development under a project that lasts for a couple of months, given that a significant positive effect from it will be evident only on the year+ horizon. If you have been developing your software for several years, automated testing can make your life much easier.
– Can you configure automated testing?
Our testing tools:
Continuous Integration: Jenkins + Selenium;
Mobile testing: Android Studio Testrecorder, XCode UI Automation, Fiddler
Automated UI testing: tests for Selenium Webdriver; Robot framework; Soap UI
Automated API testing: java+junit, jmeter
Load testing: jmeter, yandex.tank
Web testing: the latest versions of popular browsers; built-in developer tools
COST
Part-time work is more expensive, as the workload is not full and the professional has to switch to other projects. If you are ready to hire a team of professionals for six months or more, we can discuss a discount. If you are ready to hire a team of specialists at once for a period of six months or more, you can talk about a discount.
Our QA rates for engineers depend on the level of specialists, the average rates are: junior - 800, middle - 1000, senior - 1200 rubles. in hour. The cost of an hour is indicated on condition of full load
Full time
Part time
Our experience
Here you can find our open portfolio comprising the projects we are entitled to disclose.
Our experience
Here you can find our open portfolio comprising the projects we are entitled to disclose.
Professional Questions
Professional Questions
We use V-Model. Testing is planned and carried out simultaneously with all stages of the software life cycle. V-Model is one of the development practices that involves systematic testing of the software product during its development and testers’ involvement at all stages of the product life cycle, starting from collecting requirements and evaluating.
A test bed is a cluster of servers for backend, frontend, and DBMS. At the customer’s request, we can provide an additional test bed for testing by the customer. Releases are uploaded therein, as soon as a feature is ready, with a description of the release composition. At this test bed, testers perform only regression, smoke, and confirmation testing. Hotfixes are not uploaded. There is a separate server both for continuous integration system and automated testing and for load testing.
Load testing is performed in several stages:
At all stages of stress testing, system responses are measured. We measure system responses at all stages of stress testing. All data is logged; based on such data, graphs and a report are built which can be either detailed or general (at the customer’s request: it takes more time to write a detailed report, but it provides a detailed analysis of graphs).
- Stability testing (smooth increase and decrease in the load while maintaining the average load level within the expected value for several days). At this stage, testing is carried out according to user behavior emulation scenarios; At this stage, testing is carried out according to the scenarios of emulating user behavior.
- Step testing (a sharp increase in load in large steps to determine the critical threshold value, for several hours). At this stage, testing is carried out based on both user scenarios and direct server requests; At this stage, testing is carried out both by user scripts and by direct server requests.
- Stress testing (a sharp increase in the load up to threshold values and a smooth decrease to determine the optimal load and the system stabilization time). At this stage, scenarios are composed of requests directly to the server.
At all stages of stress testing, system responses are measured. We measure system responses at all stages of stress testing. All data is logged; based on such data, graphs and a report are built which can be either detailed or general (at the customer’s request: it takes more time to write a detailed report, but it provides a detailed analysis of graphs).
- Manual “black box” testing (UI+API) using browser plugins and extensions to debug tracking of transmitted information, zoom tools, other tools for interacting with a server via protocols with URL syntax, such as curl;
- Automated “black box” testing (UI +API);
- Continuous "White box” testing, i.e., unit self-tests and code review within the development team;
- Load testing to determine the system performance and fault tolerance, to identify bottlenecks, and to assess the ability to withstand the expected load.
We need:
Next, we perform stress testing: the load is applied as minus 1 step from the previous stage. What will this give the customer? The customer will understand the system failure dynamics and for how long the system will remain available; this is the time to employ urgent measures to restore the service.
- A list of the most popular requests for the service (such list is taken from server logs, but we can prepare it ourselves);
- Average (or expected) number of users per unit of time;
- Excluding the IP address, from which the load will be applied, from the ban list (in coordination with devops and the customer’s security service);
- A person on the customer’s side who can quickly restore the system after releasing the load.
- If the customer considers it possible, we can perform step and stress testing on production servers. How does it work? The load is applied in steps. The step height is selected based on the initial data and the testing progress. At this stage, we determine how the system copes with a sharp increase in load and how long it takes to process requests after each step. The last step is the system unavailability;
- If necessary, the expected response time for requests (otherwise, the time is simply measured and reported).
Next, we perform stress testing: the load is applied as minus 1 step from the previous stage. What will this give the customer? The customer will understand the system failure dynamics and for how long the system will remain available; this is the time to employ urgent measures to restore the service.
Order
You can leave a request for testing or ask additional questions here
Order
You can leave a request for testing or ask additional questions here
By clicking on the button, you consent to the processing of personal data and agree to the privacy policy.
Predictive analytics
Machine vision
neural networks
Our partners in the USA: