Ethan Lieske

My ​philosophy of SRE

      The great challenge facing tech organizations today is their clients' insatiable need for features and reliability. In many situations reliability is sacrificed on the altar of feature requests. While this provides short-term gain, it leads to long-term issues. Outside of the obvious impact of unstable software, feature-only development leads to talent retention issues, and, ironically, feature delivery delays. This is where an empowered SRE team can be positioned to balance the needs of the product with the needs of maintainable software and systems. A core tenet of the philosophy is that of platforms: deployments, telemetry, monitoring, logging, etc., should all be provided as a platform. Developers should not be wasting cycles re-inventing the wheel when it comes to these categories. Error budgets are another core tenet, this being less about implementing a by-the-book error budget and more about how you prioritize work. Development work should be prioritized in part by the current stability and maintainability of the service. A service that is constantly under-performing should not be receiving features until it is stabilized. Codifying error budgets into the company processes also reduces the constant struggle between tech and product organizations over the priority of work. It is important that SRE have a place at the leadership table, otherwise they quickly devolve into a legacy operations team that will struggle to retain talent and provide value to the business.