HN Explorer

Ionos had an unreported outage today

20 points by MarcellusDrum ↗ HN
The web hosting company IONOS (previously 1&1) had an outage today. I was woken by the call of one of my clients, saying that their website is down. I quickly discovered that all the websites I host there are unreachable. Trying to login to my account, it turns out that the official website ionos.com is unusably slow (like it takes 10 minutes for each page to load, and it doesn't load correctly). So I couldn't even contact support.

Checking their official status page and Twitter accounts, I couldn't find any mention of the incident.

After more than 2 hours, I was able to reach their customer support via phone call, and they informed me that they are having technical issues, and service will return shortly, which it did.

So my issue isn't that the outage happened. It is that they reported this issue no where. All their status pages showed no incidents, and that they are fully operational. So basically, I had to spend more than 2 hours answering angry calls from my clients, without being able to tell them what the problem is.

Their status page still shows no incidents, 6 hours after the outage.

https://www.ionos-status.com/

5 comments

[ 3.2 ms ] story [ 19.8 ms ] thread
Dunedan
I had the same experience when dealing with IONOS when I encountered a problem a while ago where the databases of managed websites were suddenly empty. As I couldn't determine why, I contacted the IONOS support, which after some back and forth admitted that they were having a database related incident going on and that the databases will get restored later on. Indeed, a few hours later the databases suddenly weren't empty anymore and contained the data they did contain before. During all of that their status page stayed entirely green as well.
NibLer
Do companies hide outages so that they would seem like error ones?
sleepytimetea
More like the error reporting service was down too. The status page was probably serving from a static page.
fxtentacle
I had the same experience with Heroku and Amazon EC2. It's industry practice that all of them polish their numbers.
playingalong
A bit less cynical take on this is that reliably assessing if you are in outage situation is hard. First of all not so easy to define. Secondly, hard to model in the monitoring. In classical attitude to monitoring you can always err on the side of bit too many alerts (oncalls may take a few false positives).

Having said that, AWS has surely the brain power and capacity to do that right. Maybe they just lack motivation.