We scraped VPS prices nightly for up to 196 nights. 273 plans never moved, and our first query missed a 40% increase

Description and use

A nightly VPS price scraper found 273 plans in five US clouds stayed at a single price, while six providers raised prices by swapping plan generations, a change the initial query missed.

Problem

The initial automated price‑tracking query incorrectly reported no price changes for netcup in September, missing a 40% increase on new plan generation.

Outcome

The study showed 273 VPS plans remained at a single price over 196 nights, while six providers raised prices by swapping plan generations or moving tiers, highlighting a blind spot in automated price‑change detection.

Workflow

• 1. Set up nightly scraping of provider pricing pages.
• 2. Store each plan’s price and billing term in a time‑series database.
• 3. Run a query to detect price changes per plan.
• 4. Compare query results against provider announcements and pages.
• 5. Identify cases where plan generation swaps caused missed changes.
• 6. Update documentation and alerting to account for generation swaps.

Evidence

Our first query said netcup changed no prices in September. That was wrong.

Source

On 22 September netcup replaced its G12 plans with a G12.5 generation that costs 40% more on the x86 VPS line and has half the disk, and no existing plan's price moved.

Source

Every active VPS and cloud plan at DigitalOcean, Linode, Vultr, AWS Lightsail and Hostwinds has held a single price for the whole time we have tracked it, 273 plans in all.

Source

The table shows one example plan from each, with its billing term. The sections after it explain each case, including the plan generations and rate types named here.

Source