This app looks to be incorrectly parsing Slack and Auth0 official status page and showing incidents as ongoing that are not
And those are just the 2 that I checked.
To be fair, accurately scraping and normalizing data from status pages is really hard to to do consistently (my company has a team of 5 engineers to do it and it's a lot of work).
Correlated downtime and this is a place I wouldn't actually mind a guess from AI on whether their is a common underlying cause between some of the things. I say AI because I don't really think anyone is going to keep all of the possible common dependencies of different privately hosted systems up to date, but AI could at least take an initial guess + try to find if anyone else is posting root cause theories elsewhere at the time and link to those (and a guess is fine enough).
Where does this draw data from? It's a similar visual concept to what we're doing at ThousandEyes within Internet Insights (see https://www.thousandeyes.com/outages/) however we make it fairly clear how we are making these determinations. Our data comes from billions of daily pseudonymous metrics from within synthetic tests running across thousands of agents around the world.
If you're drawing the data from a public resource like downdetector or using the sites status pages, then you may not be reflecting reality, but it should be clear what the provenance of the data is.
28 comments
[ 3.7 ms ] story [ 52.4 ms ] threadAnd those are just the 2 that I checked.
To be fair, accurately scraping and normalizing data from status pages is really hard to to do consistently (my company has a team of 5 engineers to do it and it's a lot of work).
``` if(github) return false ```
https://how.complexsystems.fail/#5
But then, the home page can be cached, and bots can be batched and nobody would ever notice the difference.
I've been building something like this for 12 years now.
One major difference is mine does not only rely on the "official" status page but also receive millions of reports from users about outages.
So your single pane of glass can show not just known outages but emerging ones that haven't been acknowledged yet by providers.
Also supports more than 8,000 services.
Does the size indicate anything?
If you're drawing the data from a public resource like downdetector or using the sites status pages, then you may not be reflecting reality, but it should be clear what the provenance of the data is.