Show alert on dataset page whenever indexing is off

[ckotwn]

http://dev.gbif.org/issues/browse/POR-2148

[simoncha]

Seems that indexing is off this week ...
Such a feature would have prevent me to bother Kyle about http://www.gbif.org/dataset/b5cdf794-8fa4-4a85-8b26-755d087bf531

[kbraak]

Every time we turn off indexing for maintenance or dedicated crawling of eBird, publishers write to helpdesk or comment in GitHub expressing confusion and frustration. Here is another recent example of such a case.

An alert explaining that indexing is turned off coupled with information about how GBIF indexes datasets is gravely needed. This will not only improve the publisher's experience on GBIF.org, it will also alleviate a lot of load on the GBIF Helpdesk. Thanks.

[dgasl]

I absolutely agree.

And it would be so useful If this info is also available through the api.

For institutions which do not have a web access to their own collections, serving through GBIF is a great solution. So we link to our datasets page and try to keep data permanently updated.

But when GBIF indexing is not working, our researchers will ask us why they can't see their latest data ... so we pass the ball to GBIF staff.
It would be great if we could instead use the api to show/hide a warning in our websites (i.e., the expected date of the next GBIF update, next to our GBIF link).

Thanks a lot

[peterdesmet]

I assume when indexing is turned off, it affects all datasets? Therefore, I would opt for a general warning banner at the top of the website if important services are down, and not just on specific dataset pages:

We are experiencing ... Check the status of our services here.

That banner should link to a http://status.gbif.org, similar to https://status.github.com/ (see #216)

[dgasl]

In the mean time, which is the best and fastest way to know when indexing is turned off?
Perhaps you could notify us via Twitter using #nextgbifindexing hashtag ?

Something like twitting "@gbif indexing is off: #nextgbifindexing 2017-04-05T12:00:00+0000"

If there are changes in schedule, you can just tweet again and update the info.
That way we could use twitter api to know, and not wasting time checking our gbif dataset pages again and again, or bothering gbif staff with emails.

By the way ... is indexing turned off right now?
Thanks

[MortenHofft]

There is now a system health page, api and an option for comms to do notifications on the site see more at the issue #216

On dataset pages, the crawling info is shown at the top (instead of the bottom) if you are a trusted contact for the dataset. Along with some more details. It also shows if crawling is off.

This will show if you are logged in and is a trusted contact. This means that it will show if your account email or Orcid is matching a contact on the dataset. Or for the installation, host, endorsing node or publisher.

[peterdesmet]

Really nice! Question: my account email is different from my dataset emails, so I would rely on my ORCID to see trusted information... but an ORCID cannot be added to a profile yet?

[MortenHofft]

Hi @peterdesmet yes ORCID can be linked, but this is only on UAT currently. Will be deployed to prod when it isn't a friday :)

[simoncha]

Crawling info seems a nice feature :)

I tried to log in with the email [email protected] which is the technical email provided in the 37 datasets https://www.gbif.org/publisher/2cd829bb-b713-433d-99cf-64bef11e5b3e are hosting, but had no success requesting a password.

Is it too early ? Or still Friday ;-) ?Should I use a different procedure than "forgot your password" ?

All the best

[MortenHofft]

Hi @simoncha
I don't see any account created with email [email protected] ? Are you sure you have registered?
So yes I would try a different procedure and create an account with email [email protected]

strike that. I do see it. You don't get a mail with a reset link? That is a bug then

[MortenHofft]

@simoncha The systems shows as having sent a reset link to the mail. Let me know if you still experience issues. If so I will create a bug report.

[peterdesmet]

I'm logged in, I have connected my ORCID to my account and I am viewing a dataset with my ORCD (this one)... but I don't see the extra crawling metrics.

Also, you mention:

On dataset pages, the crawling info is shown at the top (instead of the bottom) if you are a trusted contact for the dataset. Along with some more details. It also shows if crawling is off.

I don't see this crawling info at the bottom either (for other datasets or when I'm logged out).

[MortenHofft]

in that case something is radically wrong or we are misunderstanding each other.

[MortenHofft]

This is what I see at the bottom of the page when visiting your dataset without being logged in. You do not see that?

I looked over the orcid code. I had tested with wrong mock data (i don't have any datasets i can call my own) - that should be fixed now. thanks. But the other part is weird. Could you take screen shots of what you see, browsers etc. I cannot explain that. Sounds like a new bug issue.

[peterdesmet]

I’m sorry: yes, I do see that information. I was confused with “instead of the bottom”, where I only saw the citation. Didn’t notice it above the citation. My bad, no bug.