New report feature for EuroVO Registry

[Markus Demleitner]

Hi,

On Mon, Feb 20, 2017 at 03:25:23PM +0200, Kostandin Vrioni wrote:
> I would like to inform you about the new report feature of our
> EuroVO Registry. The "Registry Ecosystem health report" which is a
> public accessible report generated monthly and consist of three
> types of reports:

Great!  I like it.

Now, on to fix what troubles are left:

ROR-10, ROR-12, ROR-13, ROR-14, and ROR-15: This is a (single, known,
minor) problem on the side of the RofR -- Yulie, can you perhaps
already make promises on removing the binding of the empty namespaces
in your oai containers?

NRV-02: I'm not terribly worked up about common prefixes as
Identifiers expliticly says that authorities are opaque.  Still, I
admit these do not look like they were registered on purpose.
Theresa, Costas: could you work these out between the two of you?

HRV-10: Why are missing shortNames flagged as "anomalies"?  I admit
that short names are generally useful in displays, but since there
wasn't a consensus to make them mandatory in VOResource, we can't
really point fingers if people leave them out, can we?.  Or am I
missing something?

HRV-11: At least for ivo://org.gavo.dc/rr/q/create I can say that the
OAI interface is missing on purpose (it's our RegTAP endpoint,
and while there is an OAI interface, it doesn't have any records of
its own, which makes the RofR validator really unhappy).  These cases
will become more frequent if Registry Interfaces 1.1 becomes REC as
defined now.  So, I guess that's no longer an anomaly.

ROR-HRV-1: There's several cases here.  ivo://org.gavo.dc/rr/q/create
is discussed in HRV-11; the RofR probably can't really deal with
these just yet.  I guess fixing this could be a criterion for the
implementation of RI 1.1.
The issue with ivo://org.gavo.dc/__system__/services/registry is my
fault.  I'll look into it.

ROR-HRV-2: I don't quite understand where these come from -- Yulie,
could you comment?

SG-1: I'm surprised to see most of the services in here (chivo-lirae
is a known problem at this point, being worked on).  I wonder what
went wrong there since I can't remember when last I found a
whopping six registries down or severely broken.  What's the deeper
meaning of _int.esa.esavo.common.PersistentHTTPConnectivityException?
Perhaps there was a local issue?


Summing up: Apart from SG-1, this is all fairly benign, and I must
say for a distributed system like the Registry it looks pretty good.
In particular, it's much better than a similar report would have
looked a year ago.  And a fairly small fix at the RofR will make it
still look better.

Thanks to all who helped de-linting the Registry, and thanks to
Costas for making this public.

         -- Markus