Website Downtime Frequency and Severity Increased in Q2 2022

Uptime.com analyzes the longest and most impactful downtime periods of the quarter.

From remote work to podcasts to streaming video, consumers and businesses are more reliant on technology than ever before. However, halfway through 2022, downtime frequency is increasing, with a more broad and severe impact, according to Uptime.com, a top-rated developer of website monitoring software.

“Some of these outages stem from infrastructure issues and some from simple issues that the service providers should have anticipated and prevented,” said Mike Welsh, CEO of Uptime.com. “Regardless, the downtimes affected millions of users, and that reinforces how crucial it is for businesses to have the best in website monitoring.”

Microsoft experienced a 12-hour outage of Microsoft Azure and 365 on June 21. An unplanned power oscillation in one of the company’s data centers created delays, log-in failures and issues accessing accounts. This affected a number of Microsoft services, most notably Microsoft Teams, a significant problem for companies that use the platform for internal business communications. “This outage highlights the importance of having emergency plans in place for unpredictable outages,” Welsh said. “Businesses who rely on Microsoft Teams were likely left in the dark when their communication platform went down.”

Marketing Technology News: PolySwarm Launches Token Rewards For Internet Browsing

Google Cloud had perhaps the most noteworthy downtime event, an outage of three hours and 12 minutes on June 7, caused by two simultaneous submarine fiber cuts. This outage reduced capacity for telecom and technology companies in the Middle East and increased latency between their Europe and Asia regions. Once alerted, the company began redirecting traffic. “This fluke accident shows us that not all downtime is avoidable, predictable or even related to software issues,” Welsh added.

The popular music service Spotify had an eight-hour outage on May 30, preventing users from accessing podcasts on the platform. This was due to an expired security certificate in a third-party platform called Megaphone that Spotify uses to host podcasts. “This was an easily avoidable problem that would be quick to remedy if the provider was closely monitoring its service,” Welsh said.

Cloudflare experienced one of the shorter outages among major providers, but it nonetheless led to broad and severe effects. The one hour and 15 minute outage on June 21 came when Cloudflare was converting busy locations to more flexible and resilient architecture – routine and scheduled maintenance. This outage affected traffic in 19 data centers that handle a significant proportion of global traffic.

Because the Cloudflare content delivery network is used by some 7.6 million active websites, the outage was especially damaging and halted traffic on Twitter, Amazon Web Services (AWS), Discord, Shopify, Canva and other popular sites.

Atlassian, a leading provider of collaboration, development, and issue tracking software for teams, had a nine-day downtime event on April 4, the longest in its history. Users had no access to Jira, OpsGenie, Confluence, and other Atlassian Cloud services. The company cited insufficient system warnings and an internal communications gap as reasons for the outage. “This shows the importance of redundant monitoring and alerting, to act as a safety net and ensure warning signs don’t fall through the cracks,” Welsh said.

Uptime.com was itself founded after a major web outage and its creators couldn’t find an affordable and user-friendly solution, Welsh noted. Uptime.com’s solution checks site availability, monitors for other critical site problems and performance issues, and provides details for accurate root cause analysis.

Marketing Technology News: MarTech Interview With Peter Birch, MD EMEA, Hybrid Theory