Skip to content

Latest commit

 

History

History
109 lines (90 loc) · 5.8 KB

README.md

File metadata and controls

109 lines (90 loc) · 5.8 KB
Raw

Exchange rate service

Web service which exposes the current exchange rate of USD to MXN from three different sources in the same endpoint. Exchange rate sources:

Running server

  • Run server python manage.py runserver

Running unit tests

  • Run unit tests python manage.py tests

Cron jobs

In order to execute the command to request the lastest exchage rate for each service provider use the following command. Consider use a linux cron or kind of service Cronicle:

  • Diario Oficial de la Federación service provider request: python manage.py diario_oficial_provider
  • Fixer service provider request: python manage.py fixer_provider
  • Banxico service provider request: python manage.py banxico_provider

.env

  • Django secret key: DJANGO_SECRET_KEY=h+fwx4a2+3(!64jcprhlom=+!1uyqn *@agi93+5xf4a+b%p4+$
  • User rate limit configuration: RATE_LIMIT=1000/day
  • Fixer API KEY for request: FIXER_API_KEY=api_key
  • Fixer URL service: FIXER_URL=http://data.fixer.io/api/
  • Currency base for exchange rate (for free account use EUR): FIXER_BASE =USD
  • Banxico request TOKEN: BANXICO_TOKEN=token
  • Banxico service URL: BANXICO_URL=https://www.banxico.org.mx/SieAPIRest/service/v1/series/
  • MXN Banxico Serie ID: BANXICO_SERIE=SF43718
  • Diario Oficial de la Federación service URL: DIARIO_OFICIAL_URL=https://www.banxico.org.mx/tipcamb/tipCamMIAction.do
  • Table exchange rate HTML position: DIARIO_OFICIAL_TABLE_POSITION=6
  • Interest row position: DIARIO_OFICIAL_ROW_POSITION=2
  • Interest column position with exchange rate: DIARIO_OFICIAL_COLUMN=3

The High Level Problem

You are the tech lead for a new Credijusto project that collects client information to be run through risk models to determine if they are viable candidates for our loan products. The project has integrations to 3rd party systems like Buró de Crédito, where we can obtain pertinent client financial information to help Credijusto assess the client risk profile. In order to launch this project, we need a production test plan. Please outline, in as much detail as you can the types of tests you would run, how you would run these tests in a live production environment, and how you accomplish running these tests given you have live 3rd party integrations.

R: In this scenario we would use a hybrid approach that includes live testing and mock testing. Live tests would be making requests directly to 3PA, ideally you would configure an account or credentials that allow us to identify those test requests to eventually analyze them. For mock tests, we would use a proxy that emulates the contracts defined for the responses of this 3PA, so that we simulate its good behavior and specifically test our system.

An important point to determine is the priority of each of these tests, that is; If most of the problems come from 3PA we would have to give priority to the execution of live tests, considering that the execution time of this type of tests is longer given the http requests that have to wait, in addition to the fact that the 3PA could present false positives due to application load or fails.

On the other hand, if our system is the one with the highest number of failures, we would give priority to mock tests, to ensure the integrity of our project against the expected responses of 3PA. I believe that both tests are important to be covered against the disadvantages of each one: the delay and false positives in live tests and the lack of coverage of the mock against the 3PA that we can't control.

The Scenario Problem

Several services for which your team is responsible depend on another squad’s service in production, and mocks of that service in development and staging.

Your team uses the mocks they have provided in both development and staging. However, there are regular incidents in staging and production due to code changes made by that team which they have not propagated to their mocks. Your developers are frustrated and losing confidence in their coworkers. The tech lead from the other squad is an E2 developer and has a reputation for over-confidence and being difficult to work with.

How do you handle this situation?

R: First, I would speak with the tech lead using a model known as COIN, where I provide Context for the problem, mention specific Observations, emphasize the Impact on the product and / or project , and propose some ideas or Nxt steps for improvement.

Possible answers:

  • It isn't our responsibility: I would emphasize the impact of this problem and mention the benefits of collaborating to solve it. In addition, I would involve the team in failure alerts or notifications so that they can empathize with our problems in a tangible way.

  • I don't have time for the call: I would add a couple of alternatives to be able to coincide, failing that I would write an email with the details of the problem and the possible solutions, trying to find a compromise with the team.

  • Those proposals aren't good: Nothing is set in stone, I am open to listening to proposals that involve both teams in order to reach a solution.

Once both teach leads understand the impact, talk with the teams and commit then with proposals. The important thing is communication, we seek to agree on how to communicate the changes of this service to our team, it can be a format for each merge within the repository using tools such as slack, or creating a channel to notify relevant changes by message or email.