2019-08-21 TC Meeting notes

2019-08-21 TC Meeting notes

Date

Aug 21, 2019

7am PST / 4pm CEST

Meeting Link

https://zoom.us/j/994971678

Attendees

  • @Klaudia Pałkowska (Deactivated)

  • @Aleksandra Ciesińska

  • @Mateusz Kwiatkowski

  • @Mateusz Wedeł (Unlicensed)

  • @Sebastian Brudziński

  • @Paulina Buzderewicz

  • @Paweł Pinker (Unlicensed)

  • @Elias Muluneh

  • @Chongsun Ahn (Unlicensed)

  • @Wesley Brown

 

Discussion items

Time

Item

Who

Notes

Time

Item

Who

Notes

5m

Agenda and action item review

@Josh Zamor (Deactivated)

 

15m

Enabling other services to read requisitions data

@Sebastian Brudziński

  • Malawi will build a new service that needs to read all/parts of requisitions' data

  • Possible workaround: create user and assign them rights to all programs and SNs (same as reporting stack)

  • Alternative: Enable service-level tokens or API keys to read requisition data

10m

Better determine usage of BASE_URL and VIRTUAL_HOST properties

@Mateusz Kwiatkowski@Mateusz Wedeł (Unlicensed)

In Angola, we needed to change BASE_URL property to enable Let's Encrypt with nginx to "nginx-reverse-proxy". This works fine considering communication between services but BASE_URL property is also used i.e. email notifications which results in incorrect URLs send to the users.

10m

Stock Management behavior with FTAP/Orderables update

@Sebastian Brudziński

Notes

 

Requisition Data

 

New microservice to read data from Requisition data in order to send metrics/indicators to an interop layer.

Problem:  the Requisition API's don't support service level tokens or other api keys meant to pull all data from Requisition.

Use-cases:

  • Send the exact same requisition multiple times

  • 2 ways to send:

    • automated:  send at the end of the month (by program, send one program one day and another program another day)

    • manual retry: send all for a period and program to interop layer

 

Is this a change in how we get data out of services?

  • API:  Carrying out workflows

  • Data Pump: To get data out of silos

 

BASE_URL and VIRTUAL_HOST

 

(encoutered in Angola)

Wanted to use Lets Encrpyt and move TLS termination to OpenLMIS Ref Distro component.

Introduced another layer of Nginx (as a reverse proxy and TLS termination)

Change to BASE_URL (this is the most used variable) and VIRTUAL_HOST was needed.

Proposal:  Split the variables, one for services and another for use in sending links. Not proposing ATM the addition of another nginx

 

Question:

  • What's VIRTUAL_HOST used for (consul?)

  • Can we introduce a new variable for email notifications?

  • Can we split the usage of these variables?

Suggestion:

  • Understanding how these two variables are used in the code-base today is a good side-effect.

 

Stock Management behavior with FTAP/Orderables update

  • See forum post in Agenda above

  • We did face this / a similar question about the role of FTAP in Stock Mgmt in inital design.

    • Josh thinks he remembers that the line might have been to use FTAP to inform what went on in a physical inventory, but not for adjustment, issue or receive.

    • We should do some research here to see if we've documented this, and/or if the code is self-documenting enough for this

 

Action Items

 @Sebastian Brudziński: To look through Debezium, data-pumps, etc.  Project Casper is a good source, @Chongsun Ahn (Unlicensed) is a good reference.
@Mateusz Kwiatkowski and @Mateusz Wedeł (Unlicensed) To post on the forum and gather feedback
@Sebastian Brudziński to look through original stock mgmt design stories, wiki, and current code to see if it gives an answer to the proper role of FTAP in Stock Mgmt
@Sebastian Brudziński:  To bring up the topic of how a change in pack size should effect Stock Mgmt on the forum / another call

OpenLMIS: the global initiative for powerful LMIS software