General Discussion
Related: Editorials & Other Articles, Issue Forums, Alliance Forums, Region ForumsI'm giving Obama's troubleshooters credit. They did a great job rescuing healthcare.gov!
Granted, the site's still got issues, but at this point, any site of this complexity, with such numbers of users, is going to have some issues, and those issues are being managed.
From Ars Technica:
http://arstechnica.com/information-technology/2013/12/healthcare-gov-sort-of-fixed-good-enough-for-vast-majority/
As a result, HHS says that HealthCare.gov "runs smoothly for the vast majority of consumers." But the site's problems are far from being completely fixed. The underlying systems to which the site ties are still causing problems for some state health exchanges. And visitors used to the instantaneous user experience of most electronic commerce sites may still balk at waiting for an e-mail message to inform them that it's their turn to apply for health insurance.
In a press conference on December 1, Zients warned that demand may still exceed the site's capacity. "So to prepare for those times when spikes in user volume outstrip the system's expanded capacity," Zients said, "we will deploy a new queueing system to serve consumers in orderly fashion and to allow consumers to request e-mail notifications when it's a better time to come back to the site."
That's not to say there haven't been drastic improvements behind the scenes at HealthCare.gov. Just being up more often than it's down is a major achievement for the site, which, according to the report, was unavailable nearly 60 percent of the time in October. But it's clear that much of the work being done now should have been addressed well in advance of the site's launch. And the architecture of the site itselfin part dictated by the Affordable Healthcare Act and by the rules created to govern how individuals are screened to determine if they are eligible for careis the antithesis of what people have come to expect from a modern Web commerce site.
First thing they did was setting up instrumentation and monitoring - first order of business was to find exactly where the bottlenecks were, and where things were breaking. This kind of thing is VITAL with really high-volume web sites. It wasn't done before, and once it got done, finding problems and fixing them sped up considerably.
The registration server, which was running on virtual servers (with that demand? WHY?!) was now moved to a new group of dedicated servers, which work far better.
Once the troubleshooting team got in the zone, fixing of software defects and issues accelerated, with 300 fixes put in the system in the month of November (only 100 in October).
They added a bunch of new hardware - the site was sorely underpowered. 12 new database servers, expansions to the storage system for those servers, a bunch of new front-end web servers, addition of a lot of new network capacity.
When first launched, the site could only handle 1,100 users on the system simultaneously. Now that number's above 50,000, and if the system's overloaded, it has a queuing system that will send you an email reminder when the system's under less load.
Before, the site took over eight seconds to serve a web page when it got a request. Now that time's down to less than a second.
Sorry freepers. Have fun shedding tears over your whiskeys back in the Cave. The site you declared to be the Doom of Obamacare is working. Are there still issues? Sure. With any web site, there are issues, but the crisis you were so gleefully exploiting (while providing zero ideas for actually getting health care to the American people) has now been averted.
Blanket Statements
(556 posts)They just keep moving the goal posts
backscatter712
(26,355 posts)They'll be shrieking about the website when they're not shrieking about Benghazi, but nobody with more gravitas than Alex Jones will take them seriously...
mwrguy
(3,245 posts)Always
Cha
(297,323 posts)SoapBox
(18,791 posts)DallasNE
(7,403 posts)When I saw data that was supposed to be saved but wasn't, improper tracking on address information and an improper recovery point when the site crashed I said "and this is only what I am seeing" so I concluded that it was only the tip of the iceberg. There was also an issue that suggested a database design problem although some of that can be masked by simple capacity issues. On top of that there were possible problems with sensitive data not being properly encrypted. That is a lot of different things to tackle on a concurrent basis, complicating test scripts.
With all of those balls in the air at the same time I was highly doubtful they could pull it off in the promised timeframe. Someone did a good job of cracking the whip this time around.