Towards the beginning of the article you mention “At that time, TrueCar didn’t understand why continuous deployment was needed. There was complacency with the status quo.”
How was this complacency challenged, and by whom? How did you address concerns and fears from people who were resistant, doubtful or dismissive?
fantastic. I look forward to the rest of the articles.
Where i work, lots of different groups with different experience levels and abilities. I often hear that cant work here, were too different, we don't need it, its too hard, etc. I love reading success stories that talk about what/how they did it. I use them as reference points that i show folks.
@philcollect Even at our scale, we currently don’t have an issue with running out of TCP connection handlers on our NGINX hosts, but using NGINX Plus for the dynamic resolution definitely works for those running into issues with connection handling.
Just keep in mind, this will create a fresh TCP connection to your upstream for EVERY request, which is inefficient and will use up a lot of TCP connection handlers on the machine. In free NGINX you cannot use this with an upstream to take advantage of keepalives and connection reuse. In our situation we resolved this by purchasing NGINX Plus which…