|
Circle CI has been steadily taking marketshare away from Travis for many, many years. And in my personal experience, mostly because of Travis' slow pace of improvement. Using containers for testing blows away performance on Travis for comparable tasks. Travis had made some improvements this last year to their workflows, configuration, and platform, but too little too late. My experience in dealing with their customer service (we did have a paid plan for some time) and customer feedback (feature requests, pleas for fixes, etc) was also quite poor. I'd moved to Circle CI two years ago, and the only tasks/projects of mine still running on Travis are those which are deprecated, in suspended animation, or abandoned. For myself and my immediate peers, Travis is obsolete, and they did it to themselves. With Azure pipelines now a thing (and also far superior to Travis) I see another slow, slow death of a pioneering service. |
This is sad, but the conclusion I have come to over the last 3-4 years as well. It was anybody's market 4 years ago, but through a sustained pace I think CircleCI have out-executed Travis consistently.
The thing I find interesting is that, looking at Semaphore CI who used to be far behind Travis, they are now biting at the heels of CircleCI with their 2.0 iteration. From what I understand they are not venture backed and are a smaller team, and yet their product is starting to look like a real competitor. To me this says that Travis just failed to execute well enough, rather than Circle having more money (although that helps).