[Feedhenry-raincatcher] Demo Apps Improvement Suggestion - Get Users From MongoDB Instead of JSON Files

[Wojciech Trocki]

Thank you so much for sending this!

+100 for this functionality. That will be really a game changer for
performance tests and will help community to provide some more production
ready user store.
Demo user store should not be used for performance tests - it will produce
improper results etc.

Json file was added as example but I think it will be more beneficial to
also have MongoDB example.


*Couple notes from myself:*
I personally think we should leave json file implementation to be used by
default in the demo.
Reasons:
  + Using this by default may lead to misunderstandings that only mongo is
provided for users.
  + We have this documented and aligned with another provider which is
keycloak.

Note:  Additional work will be required to save data to mongo and provide
indexes for the queries.
We should assume that users will add database indexes for user identifiers
which is recommended in documentation.
This will have significant impact on performance.

 Actual provider implementation should be really simple - we just need to
read users from mongo.
You can see existing code here:
https://github.com/feedhenry-raincatcher/raincatcher-core/
blob/master/demo/server/src/modules/passport-auth/DemoUserRepository.ts
Best to create separate class like MongoRepository.

I can help you with contribution or even do that for you if needed.


WDYT?

On Wed, Oct 18, 2017 at 4:44 PM, Thomas Nolan <thnolan at redhat.com> wrote:

> Hi guys,
>
> I have a suggestion regarding the the population of users in our demo
> apps.
>
> I believe it would be more beneficial to read users from a MongoDB
> instance rather than the JSON files as currently happens for the following
> reasons:
>
>
>    - This will be closer to a production like system.
>    - Currently to generate a set of users for performance testing, the
>    generation tools need to generate the users, copy the output files to the
>    correct folders inside the raincatcher-core server and start the server.
>    Our generation tools could easily interact with the MongoDB instance and
>    create the user information in the database, and avoid time lost in copying
>    the files and possible errors in copying the files to the wrong location.
>    This may have the biggest impact on an OpenShift instance where it would be
>    easier to generate users by interacting with the MongoDB endpoint rather
>    than either build a new image, or change the contents of a PV, in order to
>    get the new users JSON files into a container each time we want to run a
>    performance test with a different amount of users.
>
>
> Regards,
>
> Tommy
>
> --
> Thomas Nolan
> Associate Quality Engineer
> Red Hat
> Waterford
>
> _______________________________________________
> Feedhenry-raincatcher mailing list
> Feedhenry-raincatcher at redhat.com
> https://www.redhat.com/mailman/listinfo/feedhenry-raincatcher
>
>


-- 

WOJCIECH TROCKI

Red Hat Mobile <https://www.redhat.com/>

IM: wtrocki
<https://red.ht/sig>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listman.redhat.com/archives/feedhenry-raincatcher/attachments/20171018/52c93e3e/attachment.htm>