poll for IT people about performance of healthcare.gov

[steve2470]

Being a computer nerd, I get into this stuff

6 votes, 0 passes | Time left: Unlimited
	Healthcare.gov is performing as I expected
	3 (50%)
	Healthcare.gov is performing better than I expected
	1 (17%)
	Healthcare.gov is performing worse than I expected
	2 (33%)
	Other (see my post)
	0 (0%)
	0 DU members did not wish to select any of the options provided.
Show usernames Disclaimer: This is an Internet poll

[steve2470]

1. bump nt

[steve2470]

10. humpty bump nt

[steve2470]

24. last bump nt

[RoccoR5955]

2. The NY State site

is swamped! 2 million logins in the first 2 hours.
They are admitting that things will be slow for a while.

[steve2470]

5. 2M logins, amazing..... I'm sure the main healthcare.gov site is even more swamped nt

[srican69]

3. If I were the program manager .. I would have had a phased roll out ...

If a system is designed for a certain peak load .. then it is my responsibility to ensure that

a) peak load is realistic

b) if not, either add servers or manage traffic

---

Apart from performance issues - there were functional issues too -

1) Drop boxes weren't getting populated
2) Data wasn't being persisted ( I had to reenter data after supposed having 'saved' it)
3) Data was wrongly persisted ( My wife's SSN showed up in my daughter's SSN box) when I came back to the site...

and this is just ME ( one data point) ... now imagine this with 2.8 million users .... I bet there are going to weeks of ALL-nighters with IT folks fighting post-production fires

so there you go ...

[steve2470]

4. is load balancing expensive ? I have been assured that it is not difficult to implement nt

[srican69]

6. what do you when all your servers are under full throttle? .... load balancing only applies

when you have spare capacity.

[steve2470]

8. ah ok, so they need to add a lot more servers, correct ? nt

[Ohio Joe]

13. Probably not

Most website will set up servers based on expected average traffic and perhaps some more for first opening when large loads are expected. Traffic will decline over the coming days so there is probably little point in adding more. Overall, issues have been minor and mainly due to either extremely high volumes of traffic or in some cases some back end calculation issues. None of it is a huge deal though it does seem some here are trying to make it one.

[steve2470]

14. thank you for your informative input ! nt

[Ohio Joe]

15. To get an idea of how well it actually is going...

Take a look at what faux, the one who will try to put the issues out in the worst way they can, has to say... They actually have a 'glitch watch':

http://www.foxnews.com/politics/2013/09/30/obamacare-glitch-watch-delay-in-launch-spanish-language-sites/

Not a whole hell of a lot there... For a site this large and with this many people on day 1... I call it a major success.

[steve2470]

18. I'm glad we have your professional perspective here

Even though I'm not an IT-trained person, I have over 17 years of experience with computers as a consumer. I was pretty confident the very first day that there would be stories of a huge load and some minor problems, at least.

I'm glad you can say that. We Dems need that healthy overall perspective, and so does the rest of the country.

[TheGunslinger]

22. They should have used Microsoft's Azure VM servers

Scale up when needed...scale back down when not.

[Trekologer]

23. The front end webservers are likely not the bottleneck

The www.healthcare.gov website is using Akamai's content distribution network so that likely can handle the load. It also resolves to both IPv4 and IPv6 addresses, so that made me smile a bit. Anyway, without knowing the software, architecture, or network, the bottleneck is probably something like a database. That's not exactly easy or cost-effective to simply throw more more hardware at, especially to only handle opening-day load.

[Ohio Joe]

12. "load balancing only applies when you have spare capacity."... No

Load balancing is when there are multiple servers under the hood of a website that distribute the load equally based on traffic. It is also the testing that is done to set up what the capacity for each server should be at optimum.

[Xyzse]

7. You kidding? I thought it would be crashing every which way today.

So the fact that people are being able to look in to things is already a plus.

Then again, that is such a low bar, so I can't say much.

[denverbill]

9. I can't say I'm at all surprised.

Reminds me of when the Rockies made the playoffs a few years ago and they sold tickets online. They claimed they were all set to go and it was a nightmare. Nobody I knew was able to connect despite trying for hours. Servers crashed. It was a huge mess.

I was really tempted to go to the site myself just to investigate, but didn't since I have insurance. I'm betting millions of other people DID go, just out of curiosity, or in some cases, maliciousness. A lot of that will die off tomorrow and over the next couple weeks.

Considering there are states with their own systems in addition to the federal site, I expected a cluster today. My guess is 90% of the problems will be gone by the end of the week and it will be working fine for the most part.

[Ohio Joe]

11. I think that overall, I called it pretty close yesterday...

http://www.democraticunderground.com/?com=view_post&forum=1002&pid=3748270

I've seen reports of a few minor crashes but nothing major... Some issues back end in calculations (this is done on mainframe) and some delays where response is overwhelming... Not bad at all

[Roland99]

16. Despite the problems I'm having, personally, something *did* impress me very much...

The "holding page" that keeps people from the actual account creation page.

I'm guessing (and it's been about a year since I've done any web development) they're measuring the load on the web farm and keeping people from moving on until the load hits a certain level.

[steve2470]

17. yea I thought that was smart also nt

[MH1]

19. Other ... I have healthcare from my employer and don't need to be hitting that site

when lots of people who need it might be trying to use it.

[SomethingFishy]

20. No worse than rollouts for any major game or MMO..

Would it have been great if it had been a flawless day? Sure. But shit happens.

[Enrique]

21. the NSA is no doubt laughing their asses off

they've got their Star Trek control room with doors that go "whoosh" and the HHS has jack shit.