This summary covers an interview on the Jukebox Podcast (WP Tavern) with Saumya Majumder, lead software engineer at BigScoots. The conversation explains a recent major Cloudflare outage, why CDN-level page caching matters, and how BigScoots integrates Cloudflare Enterprise to deliver fast, resilient WordPress hosting.
Who Saumya Majumder is
Saumya leads high-performance WordPress engineering at BigScoots. He builds caching systems, migration tools, worker-based automations, and edge computing solutions, and he supports enterprise accounts, internal WordPress projects, and plugins. His background includes work on projects like Super Page Cache for Cloudflare, and he focuses on developer-friendly, scalable hosting.
Why CDN-level page caching matters
Saumya describes the difference between server-side and CDN-level page caching. Traditional server caches keep HTML on the origin server, which reduces origin compute but still requires requests to travel to that origin, adding latency for geographically distant users. Many CDNs historically only cached static assets like CSS, JS, and images, so HTML still came from the origin. CDN-level page caching stores full HTML on Points of Presence (PoPs) worldwide, letting users fetch complete pages from a nearby PoP and dramatically reducing latency.
What happened during the Cloudflare outage and why failures happen
Modern web platforms are complex and composed of many interdependent systems. Cloudflare is more than a CDN; it runs many services across a global footprint. In this incident, an unexpected file size change combined with propagation behavior made the problem resemble a DDoS attack, initially misleading responders. Stabilizing global control planes and edge propagation can be slow because changes must be carefully rolled out across many regions.
Saumya emphasizes that no platform is infallible: major providers and private data centers can all experience outages. Outages carry reputational and financial costs, since enterprise SLAs often require credits or refunds. He praises Cloudflare for publishing detailed postmortems and using lessons learned to harden systems, though tiny edge-case probabilities can still produce failures.
How BigScoots reduced customer impact
For sites hosted by BigScoots but proxied through Cloudflare, the team used Cloudflare’s API to temporarily disable proxying, sending traffic directly to BigScoots origins until Cloudflare stabilized. Because Cloudflare’s control plane (API) remained accessible, BigScoots automated rapid failovers and reduced downtime. This approach doesn’t help sites running entirely inside Cloudflare Workers or Cloudflare’s hosting, as those depend wholly on Cloudflare infrastructure.
Cloudflare Enterprise features and tiered caching
Saumya explains Cloudflare Enterprise enhancements: higher cache hit ratios, tiered caching, and regional tiers. Tiered caching groups PoPs into lower and upper tiers; a lower-tier PoP queries an upper-tier PoP before hitting origin. The upper tier pulls from origin and serves multiple lower-tier PoPs, using Cloudflare’s private backbone to speed responses and reduce origin load. This design increases cache persistence globally and reduces redundant origin requests.
BigScoots’ integration and private peering
BigScoots adopted Cloudflare Enterprise early and built tight integration. Their BigScoots Cache plugin orchestrates Cloudflare page caching and provides granular controls. In addition, BigScoots maintains private links and dedicated fiber between their data centers and Cloudflare, so when Cloudflare fetches content it travels over a private high-bandwidth path instead of the public internet. Private peering improves speed, reliability, and consistency of origin fetches.
BigScoots Cache plugin capabilities
The BigScoots Cache plugin runs on WordPress sites to manage Cloudflare page caching and to add features:
– Fine-grained TTL controls and hooks so developers can set cache lifetimes per page type.
– Intelligent purging that clears related pages (taxonomies, archives, author pages) when content changes to keep caches coherent.
– REST APIs and hooks that let external systems and developers programmatically clear or manage cache.
– Custom rules and managed assistance for sites with unusual archive structures or complex theme behavior.
Portal, enterprise controls, and managed services
BigScoots’ portal gives enterprise users detailed Cloudflare controls in a simple interface: toggles for login protections and hardening, image optimization, Rocket Loader, and bot or AI crawler blocking or challenging. Geo controls allow blocking or challenging traffic by country or continent. While these settings are accessible to nontechnical users, APIs and hooks enable deeper automation for developers and agencies.
BigScoots also provides hands-on migration, performance tuning, and engineering services. They handle zero-downtime migrations, managed onboarding, and custom development when needed, while the portal supports self-service for many common tasks.
Who benefits
Although agencies and enterprise customers gain the most from advanced controls and private peering, many smaller businesses also use features like country blocking or analytics fixes. BigScoots aims to give sensible defaults plus powerful tools so a wide range of customers can improve performance and security.
Final takeaway
Outages are an inherent risk in a complex internet. Responsible providers document incidents, learn, and harden systems. Cloudflare’s scale and transparent reporting help, while BigScoots combines Cloudflare Enterprise, proprietary caching, private peering, and managed services to deliver fast and resilient WordPress hosting and to reduce the effect of upstream incidents for their customers.
For the full discussion and links, see the WP Tavern podcast at wptavern.com/podcast.