How are status pages for major CDNs and major backbone providers designed to be up even though the provider is down?
In: Technology
Popular status pages providers usually have fallbacks in case something fails. They use different DNS providers, CDNs, and hosting. Pages are often static and light, which helps a lot when dealing with a lot of traffic.
Let’s say you want to create a service on top of Amazon Web Services. You can use Route 53 for DNS and Cloudflare/Akamai/Google/self-host as fallback. You can use CloudFront as the main CDN and then Fastly/Akamai/Cloudflare/etc as fallback (or bypass the CDN). For hosting, maybe Amazon EC2 as main and Google Cloud/Azure/Linode/Digital Ocean/OVH/etc with the option to quickly scale if needed. If there’s a need to store heavy files, you can use Amazon S3 but also Akamai/Backblaze B2/Digital Ocean Spaces/etc.
To make things even more robust, you shouldn’t rely on only one location either. If you’re hosting something in US-west, have fallbacks in US-east. If possible, use different continents.
Services sometimes stop working even with all this redundancy because there’s always something that can go wrong: a bottleneck somewhere, something that fails to redirect traffic, a problem that no one thought about, one of the services use the same upstream provider, etc.
If you are a website/service operator, want to control your own status page and really want for it to be online when you’re having issues, you should at least use different services for everything (maybe a different domain and different DNS servers, CDN and hosting provider).
Latest Answers