> I'm not personally using this but people I trust recommend it highly. > It has a free option plus fairly reasonable monthly options. > > http://www.pingdom.com/
We use a similar service: cloudkick We have some custom plugins that we've written for it (to monitor RabbitMQ consumers and queue sizes -- we have had consumers die and need to be restarted). As you are, we are using Nginx. We have it set up to serve static content and act as a load balancer and to do SSL termination. We have a cluster of app servers set up behind this Nginx load balancer. Having multiple app servers helps significantly with issues like a single box going out. We have 3 configurations for the load balancer: an A, a B and A+B. Where each has 50% of the app servers configured. Nginx handles any one of the app servers being down, and we perform releases, we split the cluster in to A, deploy to the boxes in B, wait for the JVMs to warm up, the flip A and B and release to the A partition. When those JVMs have warmed up we put them all back into rotation (A+B). We use a symlink to point to the 3 configurations and do a HUP on Nginx to have it reload its configuration without dropping any connections. We're not on AWS yet, if we were I think the same would be accomplished with ELB instead of Nginx. I think this is a pretty common set up -- we are following the books: Scalable Internet Architectures, and Release IT (I highly recommend both of those books). Regards, Kyle -- Twitter: @kyleburton Blog: http://asymmetrical-view.com/ Fun: http://snapclean.me/ -- You received this message because you are subscribed to the Google Groups "Clojure" group. To post to this group, send email to clojure@googlegroups.com Note that posts from new members are moderated - please be patient with your first post. To unsubscribe from this group, send email to clojure+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/clojure?hl=en
