Search Mailing List Archives


Limit search to: Subject & Body Subject Author
Sort by: Reverse Sort
Limit to: All This Week Last Week This Month Last Month
Select Date Range     through    

[bioontology-support] [BioPortal] Feedback from tomc

John Graybeal jgraybeal at stanford.edu
Fri Oct 27 19:15:25 PDT 2017


Tom,

I never answered your question 3.5, I just realized.

3.5) Is your server often overloaded? If so could we explore another mechanism to determine if a namespace/prefix is registered/taken in your system

As other people have raised similar questions, I thought I'd offer some information.

I think our metrics are pretty accurate about how often our server is totally overloaded, although I'm not 100% sure whether the REST API metrics are best indicators of the frequency of a 500 error.

Since 8/20, the weekly UI uptime metrics are
  100.0% 100.0% 100.0% 100.0% 99.594% 98.452% 98.35% 100.0% 90.504% 84.085%

and the weekly REST API uptime metrics are
  100.0% 100.0% 100.0% 100.0% 99.597% 98.448% 99.783% 100.0% 99.938% 99.828%

Typically down times occur more often in the evenings, when many systems and robots are running large batch jobs, and no one is awake to kick the system.

Individual services may be down at different frequencies. For these numbers I used our service report from New Relic, which is doing a number of fundamental "is the system up" checks, but may not notice if a particular endpoint is down.

Whereas a bit over a year ago we were having a lot of down times from specific query types, we addressed a number of those problematic cases and the system became more stable. But more recently another problem has surfaced with corrupt data bringing the system down on occasion -- we haven't had the opportunity to thoroughly investigate this yet, and the system has behaved this week.

I would be happy to explore alternative mechanisms as you suggest—please let me know offline what kinds of things you are thinking about.

John



On Aug 6, 2017, at 5:28 PM, tomc <tomc at cs.uoregon.edu<mailto:tomc at cs.uoregon.edu>> wrote:

1) sounds good

2) the 'create an account' page and the 'provide feedback' page
from Firefox on linux with
only whitelisted 3rd party js
https_everwhere
privacy_badger
2.5) I needed to use Croinum to create an account then feedback did not present a recaptcha if I was logged in.

3) You are right, must have been unfortunate timings on my part
  but no browser caches were involved just http HEAD from a script.

3.5) Is your server often overloaded? If so could we explore another mechanism to determine if a namespace/prefix is registered/taken in your system.





   ... nice weather eh?          tomc at cs.uoregon.edu<mailto:tomc at cs.uoregon.edu>

On 08/06/2017 04:31 PM, John Graybeal wrote:
Tom,
Thank you for your input, appreciate your taking the time.
1) Yes!. It is a good point that we should make the site exclusively https, and I see there are indeed places that really need it. We'll take care of this, thanks for pointing it out. (I know in one or two cases we came across a library that isn't compatible, but we need to see if we can carve out an exception for those cases, if any remain.)
2) Where are you seeing the recapture that breaks? I was just able to successfully submit feedback with and without https protocol, at least from Chrome.
3) We agree, and BioPortal is working correctly in this particular case. Under normal operations, the portal reports "page not found" to a URL like http://bioportal.bioontology.org/ontologies/BUNNYFOOFOO. When you are seeing a 500 error, that is an actual server error, for example because too many requests are pending. Often performing a reload (sometimes after clearing your cache) will fix this.  While we have seen other cases where we are returning 500 errors inappropriately, I believe we have fixed all that we've seen. We'll continue to watch for those specific situations.
John
On Aug 4, 2017, at 1:30 PM, support at bioontology.org<mailto:support at bioontology.org> <mailto:support at bioontology.org> wrote:

Name: tomc

Email: Tom.Conlin at gmail.com<mailto:Tom.Conlin at gmail.com> <mailto:Tom.Conlin at gmail.com>

Location: http%3A%2F%2Fbioportal.bioontology.org<http://2Fbioportal.bioontology.org> <http://2Fbioportal.bioontology.org>%2Fmappings


*Feedback:*

Hi
First impressions

1) You are not using httpS even where you really need to.
see letsencrypt/certbot

2) The recaptcha is broken in secure browsers so one
can't create account or feedback (possibly related to #1)
I had to create this account with a browser I never use

3) in http "page not found" is a 400 level error not a 500 (server error)
i.e
http://bioportal.bioontology.org/ontologies/BUNNYFOOFOO

thanks and sorry to be terse, must keep moving


_______________________________________________
bioontology-support mailing list
bioontology-support at lists.stanford.edu <mailto:bioontology-support at lists.stanford.edu>
https://mailman.stanford.edu/mailman/listinfo/bioontology-support
========================
John Graybeal
Technical Program Manager
Center for Expanded Data Annotation and Retrieval /+/ NCBO BioPortal
Stanford Center for Biomedical Informatics Research

========================
John Graybeal
Technical Program Manager
Center for Expanded Data Annotation and Retrieval /+/ NCBO BioPortal
Stanford Center for Biomedical Informatics Research
650-736-1632


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.stanford.edu/pipermail/bioontology-support/attachments/20171028/bf5a4a9e/attachment-0001.html>


More information about the bioontology-support mailing list