Case Study: Cloud DevOps vs. Ticketed KPI Ops

Quick analysis: I got stuff done.

Details:

Last week we identified a need for some functionality on the website which would best be served by a search engine (as opposed to a traditional RDBMS).

I went online to our cloud services provider (Heroku), and found a viable and financially sensible option (Bonsai ElasticSearch). Within a couple of hours, we had an instance up and running with our development data. Within a couple of days, we had heterogeneous documents indexed on all fields, and a geospatial search feature for deployed IoT devices.

Compare to a relatively recent job in which I had to submit a ticket, wait six months for a VM with inadequate permission, and then choose between letting the feature request die on the vine, or utilizing a text indexing feature in an RDBMS that was already overloaded and performing poorly.

Takeaways:

1. If you don't give your development team the infrastructure they need, your solutions will suffer. Put up walls, you will find they prevent motion. Ops managers love tickets and KPIs and thus fail to meet the needs of technology solution delivery. Stop the make-work mentality and cooperate with flexible solution delivery.
2. I am an experienced senior architect who has worked on search engine integration before. A greenhorn would not be able to turn around the infrastructure effectively. Cloud services do not replace experience, as much as some managers want to believe it. Stop the madness of turning over architecture to junior devs who have an AWS account.

Successful delivery of technology solutions is contingent upon putting competent people in a position to work efficiently. Your competitors will thank you for failing to do so.