On 27/02/2020 07:59, Martin Grigorov wrote: > Hi, > > On Wed, Feb 26, 2020 at 5:40 PM Mark Thomas <ma...@apache.org > <mailto:ma...@apache.org>> wrote: > > Hi all, > > I took a look a the download stats from downloads.apache.org > <http://downloads.apache.org>. > > I didn't look at what was being downloaded (hashes, signatures or > releases). I just compared number of requests for each version over a 24 > hour period (yesterday). The results were: > > Tomcat 3 0.01% > Tomcat 4 0.03% > Tomcat 5 0.11% > Tomcat 6 0.39% > > > > Tomcat 7 26.5% > > > The percentage for Tomcat 7 is rather high at the moment. > If the number is similar for the regular runs, could it affect our > desire to EOL it next year ?
I think it is a bit chicken and egg. Although it is relatively low risk to upgrade from one major Tomcat version to another, there are a reasonable number of users who will stay on their current version until EOL is announced. Previous major Tomcat versions have had a life-time of around a decade and Tomcat 7 is going to about the same. It just occurred to me that downloads.a.o is only going to show current versions. The equivalent figures for archive.a.o (same day) are: Tomcat 3 0.23% Tomcat 4 0.30% Tomcat 5 1.43% Tomcat 6 7.12% Tomcat 7 28.1% Tomcat 8 46.7% Tomcat 9 15.9% Tomcat 10 0.18% Committers can also get download stats for Maven Central by logging on to repository.apache.org There are some "interesting" numbers there. It looks like the results are significantly skewed by something (CI systems?) repeatedly downloading the same old version. I do wonder how many of the downloads.a.o and archive.a.o requests are also from CI and similar systems. Mark --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@tomcat.apache.org For additional commands, e-mail: dev-h...@tomcat.apache.org
