The production environment
Posted: Sun Dec 22, 2024 8:24 am
This simulates the load coming from multiple concurrent clients, each traversing the list of URLs in its own random order, to create some unpredictability. Once all threads have completed, run_test()shows the average request time for each request URL, as well as an average of all requests combined, which is the metric I decided to use for my analysis. The command line arguments allow me to pass the server root, API key, request start and end dates, and concurrency. With these commands, I can test different scenarios. The test script is now ready, so it's time to get some indicators! Testing on the development system The development system I'm working on is a Mac laptop with 6 hyperthreaded cores and 16GB of RAM.
The production environment for this dashboard is a Linode virtual server whatsapp philippines number with 1 vCPU and 2GB of RAM. From my past benchmarking experiences, I know that the results of a fast system are not always the same as those of a slower system. So my ultimate goal is to test the production system and make decisions based on the results obtained on that platform. But before that, I wanted to run a first series of practical tests on my laptop to make sure that the test script worked correctly, but also because I was curious to see how these two databases behaved on a fairly powerful platform.
The testing methodology I decided to use is as follows: test the system running under both databases, for queries with periods of one week, one month, one quarter and one year, with all queries having 01/01/2021 as the start date; repeat the tests with 1, 2 and 4 concurrent clients; for each individual test, run the script three times and record the best of the three; and use the total average of all queries as a metric. With this plan, I achieved 24 data points (2 databases x 4 query periods x 3 concurrency levels). The following graph shows the response time for PostgreSQL (blue) and SQLite (red), with a single client.
The production environment for this dashboard is a Linode virtual server whatsapp philippines number with 1 vCPU and 2GB of RAM. From my past benchmarking experiences, I know that the results of a fast system are not always the same as those of a slower system. So my ultimate goal is to test the production system and make decisions based on the results obtained on that platform. But before that, I wanted to run a first series of practical tests on my laptop to make sure that the test script worked correctly, but also because I was curious to see how these two databases behaved on a fairly powerful platform.
The testing methodology I decided to use is as follows: test the system running under both databases, for queries with periods of one week, one month, one quarter and one year, with all queries having 01/01/2021 as the start date; repeat the tests with 1, 2 and 4 concurrent clients; for each individual test, run the script three times and record the best of the three; and use the total average of all queries as a metric. With this plan, I achieved 24 data points (2 databases x 4 query periods x 3 concurrency levels). The following graph shows the response time for PostgreSQL (blue) and SQLite (red), with a single client.