Better Dev Link

It covers log, metric, healthchecks, alerting, visualization with grafana, custom miniprofiler and client timing(which is usually ignore because we only measure server timing)

Dan works on React, he creates this site to share many article about different aspect of React. Simple thing that we overlook, which turns out more complex than they are.

Rust is about to hash a faster Hashmap implementation, based on Google’s SwissTable implementation. This post talk about many thing about Hashing and Hashmap: Dealing with collisions, CPU Cache. You should still read even if you don’t write Rust

Betterment write a custom CLI to generate CircleCI config, based on a certain rule to standard the interface such as project type, repo layout, build script. The custom CLI use a DSL in YAML, which make it re-useable acrorss teams.

Writing correct concurrent application is never easy as you have to deal with several concurrency issues like race conditions, livelock, deadlock, memory interference. We will approach this problem as an interview question.

We can run multi processes, multi user, multi tenants in same computer. How we can protect these? How we can say these process are together and they cannot see other process?

Ever feel bad if you think Kafka is complex. Even NewRelic didn’t understand all of the key indicators to monitor and alert on. But you will know after reading this. Also check out How Pinterest runs Kafka at scale

Take a step back and think about these HTTPS concept/infrastructure in real world. Is the private key really that protected? Or it sit in your laptop right now open to all that NPM malware. Same thing to CA private key. Or we maynot even trust the CA anymore, example Symantec’s CA incident recently.

HTTP/2 is complex, in a single TCP connection, it transform packets of multiple requests-response, in either direction. To make it worse, it even support priority with weight. This complicated the watefall you see in network request timeline.

The benefits of a duty-segregated path to production are: 1. Debugging faster and more easily by isolating development and testing into stages. You can detect problems earlier, and feedback loops are faster. 2. Saving resources by keeping the number of production-like environments as low as possible (one or two is good).