An online store built on OT Commerce faced severe performance degradation and sharply rising operational costs when automated bots began systematically crawling product pages. CPU utilization repeatedly hit 98–100% while the number of paid calls to the external OTAPI platform increased six- to seven-fold over three days.
The owner initially described the problem as a DDoS attack because requests arrived from many different IP addresses across multiple countries. However, the traffic was not a volumetric network-layer assault but rather automated HTTP requests that triggered heavy server-side logic and billable OTAPI operations.
The store uses the ready-made OT Box solution, which integrates with the OpenTrade Commerce platform via OTAPI. When a visitor opens a category, search result or product card, the engine may call paid methods such as SearchItemsFrame, GetItemInfo or GetItemFullInfo to fetch live data from Taobao, Tmall, 1688, Alibaba and other marketplaces. Each such call is metered, and some operations carry a multiplier cost.
A previously installed anti-bot module running directly on the VPS failed to stop the problematic traffic. The module’s vendor confirmed it was not designed to address large-scale parsing scenarios. After the owner restricted access to product cards for unauthenticated users, the automated activity persisted.
Switching the site behind the CRONARMOR WAF reversed the situation. Within hours of the cut-over on 16 September, CPU load fell from peaks of 80–100% to mostly single-digit percentages. Over the following 24 hours the server rarely exceeded 9% utilization.
Filtering statistics for the partial period showed 279,680 total HTTP requests, of which 276,100 targeted web pages. Automated requests blocked at the earliest stage numbered 115,660 (41.9%), with 99.3% of them never reaching the origin. Behavioral bots visible in analytics accounted for only 1,367 requests (0.5%), while 27,190 legitimate search-engine crawlers were explicitly permitted.
The case illustrates that high CPU combined with rising external API costs can result from scraping bots rather than classic L7 DDoS. Proper distinction requires analysis of request patterns, not merely the number of source countries or visible analytics events.
