What i learned from big web app deployments
[EuroPython 2012] Christian Theune - 3 JULY 2012 in "Track Lasagne"
The requirements and expectations towards deployment and operations are advancing at an incredible pace. Especially the DevOps community encourages and demonstrates how system administrators can leverage methods from other areas of expertise to achieve classical goals: fewer outages, more flexibility, lower cost, higher security standards, improved reproducability and overall increased quality. The hosting of complex, traffic-intensive web applications nowadays requires a solid and flexible infrastructure: you need to take care of all the homework first before going into the nitty gritty details of your specific application. My talk shall demonstrate the aspects we meet on the infrastructure and service level when deploying and operating complex services. I’ll show-case how we approach the problems we face to prepare for an ever-improving future. As many people and companies are experimenting with different approaches today, I’d love to spawn a separate discussion, maybe in an open space, to talk about why you do something different from us and what your experiences are. The topics will specifically include: infrastructure versus service, data center operations, make-or-buy, automation, virtualisation, monitoring, help desk, SLAs, deployment checklist, staging, secrets, HTTP/S termination, load balancing, reporting As you can see, the talk’s topics are mainly focused on methods and lines of thought. However, if you’re interest in tools - I will also mention the following on a varying level of detail: puppet,
haproxy, nginx, apache, varnish, supervisor, cron, Linux, DELL, ProCurve, KVM, Nagios, RT, fabric, zc.buildout, batou