In May I was able to attend DevOps Days Austin and last week I attended O’Reilly’s Velocity in Santa Clara. Anyone that is involved with hosting technology has run into the phrase/word “DevOps.” Usually it results in a little head scratching and then questions around “what is DevOps?” Good question. The honest truth is that we (as in the larger technology industry “WE”) are still trying to figure that out. Both DevOps Days and Velocity dive into the tactical aspects of DevOps but I wanted to focus on the overall thought process of DevOps.
I recently attended the CIO Europe Summit, hosted by CDM Media. There was a first for me at this event: a CTO for a major global bank gave a 30-minute talk on DevOps to his peers. (In fact, he credited Rackspace’s awesome YouTube Video as a great starting point).
Rackspace and our team of DevOps Automation specialists are headed to San Francisco next week for ChefConf 2014! Be sure to drop by booth #126 to find out more about our DevOps Automation Service and managed support for DevOps tools to help automate the process of deploying and scaling applications.
It was only a few years ago that Rackspace did everything manually to launch new products. The process would take us two years and by the time the product launched, someone would beat us on time to market and it would diminish our hard work. We were entering the cloud market and it was clear that we either had to innovate more quickly and get products out faster or we would not be competitive in this growing, fast-paced market.
Maintaining host files on standard *nix system has been traditionally done by hand. This becomes a challenge as the number of systems grows and is more true in the cloud model, where you might add/delete servers at a higher rate. One solution would be to use DNS and use a local zone to store your host name to IP mapping. If you are using Chef for automation, here is another example on how to automatically generate the host file entries.
Alerting is so passé. Think about it. How much of an outage is spent on alerting? Two minutes, maybe 15 if you’re unlucky and the person you’re looking for is not-so-easy-to-find. There are lots of services that alert you when something broken, but what happens after that?
For the fast-growing Internet-centric and SaaS applications that you deploy weekly or even daily, you need specialized DevOps engineers who can look beyond servers to help you build the most agile and scalable platform possible.
You are pouring a sizeable portion of your life into building a world-class product. Amid the stress and passion of building out new features and fixes, it is important to also develop the discipline of automated testing and configuration management. Here’s why: