- how to get a startup going on the side
- lessons learned from when things go wrong and servers crashed
- smoke tests for your application
- does that app work even those the site is up
- start simple
- set hard deadline and release on that date
- focus on core features
- engage with customers as much as possible
- don't implement everything that people want, choose the highest value features
- User Experience Alerts
- Is the app down or query taking too long to load
- System opperation alerts
- Done with cron to check server stats and alert if broken
- System Analysis Alerts
- DB logs
- Exception alerts
- Indexing
- Explain analyze is your friend in Postgres
- Concurrent Indexing
- origianlly used DelayedJob
- replaced with ZeroMQ
- Dedicated staging environment
- Application Deployment
- continuous deployment with zero downtime
- Application Configuration
- Standardization
- Look at replacing mySQL with Postgres
- Git process standardization, Git-flow http://jeffkreeftmeijer.com/2010/why-arent-you-using-git-flow/