I mean, I understand the disincentive to do this, but for something like AWS (where error icons are even more subtle than this) I care about accuracy way more than looking scary.
Red? Not bloody likely. We should be cheering Apple for even using yellow, most of the time when things are on fire we don't even get that. Instead we get "green, with the tiny blue 'i' in the corner". Status pages are a bad joke.
Yeah. Like it or not, the status page reflects on the brand identity and having a status page that reflects the real condition of the infrastructure is immensely harmful to sales. I've found that most status pages suppress information about important events more than they publish it.
I wonder if there's an "engineer's feed" out there somewhere that looks like mumbo-jumbo to the layperson but could actually be useful in allowing companies to disclose issues without the softer side of the company flipping their lids.
Separately, many of the Status pages out there setup with the Statuspage.io service also incorporate "Pingdom" like features where heartbeats and/or checks are automatically collected and notifications and updates set without intervention.
Salesforce is one of the odd-balls that is 100% honest about issues with their service. One of their sandbox instances can have minor issues for 5 minutes (not often mission critical, just performance degradations) and I get an email - even for instances I do not use. I usually get a couple emails a day, and the trust.salesforce.com site shows the full history (including a colored timeline) of the availability of each individual instance.
It's nice to see they've carried on this same policy with Heroku, for all the flaws the company has they put their money where their mouth is.
I deployed a status board once. I was forced by the higher ups to take it down. Something about it sending the wrong message to our clients. Sending out emails once an hour with status updates was way more productive.
I have the "internal status dashboard" (read: "notifications of individual instance or app downtime") for us, and the "external status dashboard" (which is the app status from the front of the load balancer) for the higher ups to show to customers.
Our network monitoring software has a dashboard we use internally. Our external monitor was a Google Apps project that was manually updated. Never got far enough into it to automate it, yet.
For all their faults (and there are many), OVH's weathermap [1] and status page [2] are immensely good and seem surprisingly honest.
On the flip side, I've had a total inability to use EE in North Yorkshire for about 2 weeks and yet their status checker reports absolutely no problems in my area (whilst Level 2 support admits they won't have an engineer on-site until next Wednesday now).
Your suggestion for an engineer feed would be greatly appreciated; there's nothing worse than spending half a day trying to narrow down an issue only to find out it's a known and unreported fault... "to err is to be human, to lie about it is ruddy infuriating".
No better than anyone else here really. The sad reality is that often these status pages are manually updated by the service teams themselves and when there's an issue you're busy trying to fix the problem, not update the status page.
Most startups / side projects these days seem to use a cloud hosting service... But you're project is meant to work especially when the cloud hosting providers are down.
Where are you hosted? Or do you host on multiple services?
FYI: "An error occurred during a connection to statusgator.com. The OCSP server experienced an internal error. Error code: SEC_ERROR_OCSP_SERVER_ERROR"
That's cool, but it'd be cooler if it did its own tests on what can be seen from the outside instead of exclusively mirroring data from corporate-run status pages, which can't be trusted.
Could be due to the limited nature of the outage. They state only some users are affected, which I can verify - App Store works fine for me, for instance.
I mean, I understand the disincentive to do this, but for something like AWS (where error icons are even more subtle than this) I care about accuracy way more than looking scary.