We’ve all faced the dilemma: should we rush a quick feature out the door, or spend extra time refactoring code? That choice often creates technical debt—shortcuts in code, architecture, or process that make future work harder. Left unaddressed, debt piles up and slows your entire team down.

When something breaks in production—maybe a server crashes or a deployment goes sideways—it’s tempting to point fingers and hunt for the person “at fault.” But that only breeds fear and discourages honest discussion. A blameless postmortem flips the script: it treats every outage as a chance to learn how our systems and processes can improve, rather than a moment for shame.

In the software development world, teams often face the dilemma of whether to build a tool in-house or purchase an existing solution. This “build vs. buy” debate can significantly impact a company’s efficiency, budget, and focus. A quintessential example of this debate is the implementation of feature flags.

The primary purpose of one-on-one meetings is to give your reports a chance to discuss issues and topics that they find important. It’s an opportunity for them to bring up concerns, challenges, and ideas that might not surface during regular team meetings. This is not a meeting to get status updates on projects! Schedule other time for that.

Wanted to share the results of the work we’ve been doing to improve performance and stabilize our ES cluster.

If you’re looking for moment.js alternatives, I’d recommend reading this article about smaller, lighter-weight moment.js alternatives.

Wanted to share some insights my team has found with changing sharding strategies for ElasticSearch (ES).

Seeing high garbage collection with ElasticSearch? My team was seeing periodic 300ms+ garbage collection pauses. We found out that we had misconfigured our ES instances.

Was doing some test speedup/performance improvement work recently on the search API and found out something; the moment.js library takes around 100 microseconds (or .1 milliseconds) to create a new instance.

My team at work recently upgraded our codebase to use to node 6.9, as node 6 has recently gone to LTS.