Blog
Practical guides on AI/LLM engineering, cybersecurity, and the developer workflows behind Rocket Dev Tools's tools.
DynamoDB RCU and WCU Explained: How AWS Actually Bills Read and Write Capacity
DynamoDB capacity units don't map cleanly to 'requests' — item size, consistency mode, and transactions all silently multiply what you're billed for. Here's the actual math.
S3 Glacier vs. Glacier Deep Archive: Which One Should You Actually Use?
Both are 'cheap cold storage,' but the retrieval cost and wait time differences between Glacier Instant Retrieval and Deep Archive can swing your total cost by 10x depending on access pattern.
AWS ALB Pricing Explained: The 4 LCU Dimensions AWS Actually Bills On
An Application Load Balancer's bill isn't just an hourly rate — it's driven by whichever of four LCU dimensions is highest. Here's how that math actually works.
Security Headers Checklist for 2026: CSP, HSTS, and the Headers That Actually Matter
A practical, non-exhaustive checklist of HTTP security headers worth setting in 2026 — what each one prevents, sane starting values, and where developers usually get CSP wrong.
Don't Paste That Into ChatGPT: A Practical Guide to Scrubbing PII and Secrets Before You Prompt
API keys, customer emails, internal hostnames — a lot of what developers paste into LLM chat windows shouldn't leave the browser. Here's what to scrub and how to do it without slowing down.
LLM Token Counting 101: Why Token Counts Quietly Control Your Cost and Context Budget
Tokens, not characters or words, are the real unit of LLM pricing and context limits. Here's how tokenization actually works, why it varies by model, and how to stop guessing.