For example, we have recently conducted a design stress test. You can answer the question about how to disable a service, or describe how a service fails. For example, we have recently conducted a design stress test.
You can answer the question about how to disable a service, or describe how a service fails.
Reply content:
For example, we have recently conducted a design stress test.
You can answer the question about how to disable a service, or describe how a service fails.
It is difficult to mount PHP, Memcached, and Redis. Some hack skills are required, which will not be used normally.
Nginx and MySQL are relatively simple. The former uses a stress testing tool similar to apacheloud to heap concurrency. If it cannot be processed, it will naturally be 502. The latter will be crazy about plug-in and plug-in, if the data in a single table is too big, it will blow up.
The purpose of the test is not to disable the service.
I personally think it is normal that testing should adopt a standard method to ensure that the application can meet the requirements.
Since it is difficult for the tester to create high concurrency that is sufficient to pose a threat, the CPU and I/O resource usage of these service-related processes will be limited.
Note: You can also migrate these services to a computer platform with low resources, such as Raspberry Pi. Who knows wwwwwww?