Hi everyone,
hope you had a good holiday break, congrats on making it to 2021!
Following on from last week’s post about the new features and product enhancements we launched in 2020, we thought we should also take a look back at how we did in terms of uptime in 2020. Customers around the world rely on our service, and - correctly - have the expectation of 100% uptime.
So how did we do in 2020?
We have many different internal system to monitor uptime and performance across our fully redundant server farms, but for this analysis we rely on an independent 3rd-party monitoring service: StatusCake. You can see our public status page, hosted by StatusCake, at status.opencagedata.com.
Based on this data we had 25 minutes of downtime in 2020, which is much worse than the 4 minutes of downtime we had in 2019. Put another way we had 99.995% uptime in 2020 and 99.999% uptime in 2019.
So, why did we do so much worse than 2019?
As you may recall, on July 17th, 2020, Cloudflare, one of the main DNS and networking providers in the world had a fairly severe outage. There is a detailed and informative explanation on the Cloudflare blog of exactly what happened, but the end result was that the internet in Europe was basically out of operation for a few hours. That single incident accounts for 15 minutes (60%) of last year’s measured downtime.
One thing it’s important to note is that downtime does not imply all services are unreachable, but rather only that some of them are, or quite simply that StatusCake’s probes are unable to reach our services. In fact during all of the measured downtime many customers continued to use our service without interruption. Whether or not a customer could reach our servers depended on their network configuration (or, more likely, that of their ISP)
While 99.995% uptime is hardly terrible for a highly affordable service like ours, rest assured that we’ll be doing our best to get back to our 2019 levels in 2021.
Happy new year and happy geocoding,