My first week of SDC was spent missing the forest for the trees. I over-invested in SQL, digging into the syntax and the tools, with the goal of coming up with really precise and performant queries.
But I felt like I was getting myself into trouble. Why should I perfect each step when I was missing the bigger picture of what the path looked like? I hadn't so much as thought about many of the concepts involved in this project - concepts like requests per second or load-balancing. Which meant that I didn't have a sense for how I should allocate my time to accord with the difficulty or depth of each step.
So I changed tact, going broad instead of going narrow. That's where I am now, with a pretty clear high-level view of SDC and the ability to prioritize and optimize in order to meet my deliverable and just to lay down some strong thought on what it will take for me to get there.
Here's what I've done:
- Moved my Express (service) server and Postgres server each to their own EC2 instance.
- Learned the dev-environment testing tool Httperf and its production-analog Loader.io
- Spun up several EC2 service instances and load-balanced them with Nginx.
Here's what's missing entirely:
- Caching. For performance enhancement.
- Integrating with the FEC front-end.
Here are some of the gaps:
- Tight, efficient queries.
- Performance testing at different endpoints (variable requests).
Top comments (0)