Working on different systems, more often than not I see that the performance capabilities of the system are either not clearly visible and defined or they are completely missing.
A lot of the developers developing features are not aware of the impact their code does to performance. By Impact I mean visible numbers of the load the system can take that go together with every build.
So the question here is:
Is performance testing your system a part of the continuous delivery pipeline or is it done as a periodical effort separated of the release process?
Top comments (2)
Periodical effort. We keep an eye on the monitoring systems, and have set up alerts for the spikes. We have a general idea of what's the current performance and the max capabilities of the current configuration.
The apps are not that big though
Speaking specifically about small companies, in my experience, performance is not something taken into account the way it should...
What I see the most is people developing things and releasing it because "it works", which leads to a bugfix 6 months later, since the aforementioned feature is now a function called hundreds of thousands of times a day, destroying the app server and the database...
On the other hand, we have larger companies used to larger-scale development and data processing, I look forward to other people's reports on that :P