‘Software glitch’ to blame for global Cloudflare outage

Approximately 10% of the internet suffered 502 errors during a chaos-ridden 30-minute outage

CloudFlare

Cloudflare has resolved an issue that led to websites serviced by the networking and internet security firm to show 502 Bad Gateway' errors en masse for half an hour yesterday.

From 2:42pm BST the networking giant suffered a massive spike in CPU utilisation to its network, which Cloudflare is blaming on bad software deployment. This affected websites hosted in territories across the entire world.

Once this faulty deployment was rolled back, its CTO John Graham-Cumming explained, service was returned to normal operation and all domains using Cloudflare returned to normal traffic levels.

Advertisement - Article continues below

"This was not an attack (as some have speculated) and we are incredibly sorry that this incident occurred," Graham-Cumming said.

"Internal teams are meeting as I write performing a full post-mortem to understand how this occurred and how we prevent this from ever occurring again."

The incident affected several massive industries, including cryptocurrency markets, with users not able to properly access exchanges like CoinMarketCap and CoinBase.

Cloudflare issued an update last night suggesting the global outage was caused by the deployment of just one misconfigured rule within the Cloudflare Web Application Firewall (WAF) during a routine deployment. The company had aimed to improve the blocking of inline JavaScript used in cyber attacks.

Advertisement
Advertisement - Article continues below

One of the rules it deployed caused CPU to spike to 100% on its machines worldwide, and subsequently led to the 502 errors seen on sites across the world. Web traffic dropped by 82% at the worst point during the outage.

"We were seeing an unprecedented CPU exhaustion event, which was novel for us as we had not experienced global CPU exhaustion before," Graham-Cumming continued.

Advertisement - Article continues below

"We make software deployments constantly across the network and have automated systems to run test suites and a procedure for deploying progressively to prevent incidents.

"Unfortunately, these WAF rules were deployed globally in one go and caused today's outage."

At 3:02pm BST the company realised what was going on and issued a global kill on the WAF Managed Rulesets which dropped CPU back to normal levels and restored traffic, before fixing the issue and re-enabling the Rulesets approximately an hour later.

Many on social media were speculating during the outage that the 502 Bad Gateway errors may be the result of a distributed denial-of-service (DDoS) attack. However, these suggestions were fairly quickly quashed and confirmed to be untrue by the firm.

"The impact of the Cloudflare outage shows the sometimes-unexpected impact of massive success - much as with early outages at AWS and other cloud providers, it's a reminder of how dependent the internet ecosystem can become on the utility and expediency of a singular platform," analyst for the cloud transformation channel with 451 Research Carl Brooks told IT Pro.

Advertisement - Article continues below

"Cloudflare has a lot going for it: it effectively ended DDOS as an attack platform as we knew it, for instance, and it's a vital performance booster for extremely reasonable prices, but it has also quietly become a part of the backbone of the internet, and like every other provider out there, it will have hiccups."

Featured Resources

Staying ahead of the game in the world of data

Create successful marketing campaigns by understanding your customers better

Download now

Remote working 2020: Advantages and challenges

Discover how to overcome remote working challenges

Download now

Keep your data available with snapshot technology

Synology’s solution to your data protection problem

Download now

After the lockdown - reinventing the way your business works

Your guide to ensuring business continuity, no matter the crisis

Download now
Advertisement
Advertisement

Most Popular

How to find RAM speed, size and type
Laptops

How to find RAM speed, size and type

3 Aug 2020
How do you build a great customer experience?
Sponsored

How do you build a great customer experience?

20 Jul 2020
Labour Party donors caught up in Blackbaud data breach
data breaches

Labour Party donors caught up in Blackbaud data breach

31 Jul 2020