Guest co-host Andrew Shafer and Coté talk with John Allspaw (now at etsy) about the dev/ops idea and, more interestingly, cultural and process needs and changes.
To listen to this podcast, download the podcast directly, subscribe to the blog/podcast feed in iTunes (or whatever), or click play below to hear it:
- Check out the video of the Velocity talk we reference frequently. It has some actual meat when it comes to biting off a dev/ops culture, practices, and process.
- John opens up by speaking to tools vs/& culture
- I ask John how his org got over the hump of “we’re a special snow-flake” – resistance to change .
- Effecting cultural change by speaking to problems solved and benefits of the new culture.
- What are some major changes in ops and dev process/culture? Continuous deployment (pushing small changes, often).
- Getting beyond the culture of no – change management tends to be really “no management.”
- You have to have a good track record to do this (MTTR, MTTD) – don’t crash the car. What’s deployment to incident ratio?
- The role of metrics (monitoring) becomes important.
- Physical vs. virtual stuff in operations, EC2 use, etc. People use public cloud stuff for best-of-breed solutions, like SmugMug usage.
- Taking flickr from 25th most popular web property to the 5th.
- Coté is glad John’s at Etsy, cause it fuels the best blog du jour, regretsy.