top of page
Search

Plan and organize the synchronization of your QE Score data collection.

Updated: 6 days ago

The "QE Score" is a living indicator: for it to accurately reflect the reality of an application, it must be supplied with fresh, up-to-date data. Collecting data on a one-off basis or too far in advance can distort the analysis and mask any drifting trends. That's why it's crucial to implement a strategy of regular, even daily, data collection for certain data sources such as SonarQube, Jira or GitLab. This regularity ensures the reliability, responsiveness and accuracy of the score, providing a real-time view of project quality. The choice of data collection method - synchronous or asynchronous - then becomes strategic.

A reliable QE Score depends on fresh data: collected regularly, in synchronous or asynchronous mode, or using a hybrid approach.

The challenges of collecting quality metrics

​

  • Disparate tools: UI tests, APIs, security, performance, code quality, etc. Metrics are scattered across different tools and platforms.

​

  • Frequency and updating of data: Some metrics are updated in real time, while others are updated less frequently, which complicates their analysis.

​

  • Complexity of integration: Retrieving and centralising this data often requires specific developments and time.

Synchronous mode: immediate and reliable, but more demanding


Synchronous mode is based on event-driven operation: each time an event occurs (e.g. a SonarQube analysis ends, a Jira ticket changes status, a commit is made to GitLab, a functional test automaton ends), the data is immediately sent to the database or Google Sheet.

in synchronous mode, the data is immediately sent to the database or Google Sheet to update the QE Score

✅ Advantages :

  • Real-time data: information is updated as soon as an event occurs.


  • Fewer unnecessary requests: no need to scrape or regularly query APIs, which reduces network load and API costs.


  • Greater reliability: the QE Score accurately reflects the current situation.


⚠️ Disadvantages :

  • More complex implementation: each tool must have a sending module (e.g. webhook, event hook, external service).


  • Tool dependency: not all tools support this type of integration without addition or adaptation.


  • Additional maintenance: if the API or event changes, the sending modules need to be updated.


Asynchronous mode: easy to deploy, but less responsive


Asynchronous mode works by periodic interrogation: a script or the Google Sheet itself retrieves data from the APIs at defined intervals (every hour, every night, etc.).


in asynchronous mode, a script retrieves data from the APIs at defined intervals to update the QE Score

✅ Advantages :

  • Easy to set up: no need to modify tools or add modules, a simple API call is all that's needed.


  • Standardised: works with all tools that have an API, without complex configuration.


  • Ideal for a first level of automation: provides relatively fresh data without any major technical effort.


⚠️ Disadvantages :

  • Potentially obsolete data: there may be a time lag between when an event occurs and when it is retrieved.


  • Regular, sometimes pointless requests: if the data has not changed, API calls are redundant and consume resources.


  • Less suitable for real-time use: a QE Score updated only at night will not reflect changes during the day.


example of a schedule for retrieving information from tool APIs
example of a schedule for retrieving information from tool APIs

A hybrid approach: combining the best of synchronous and asynchronous for your QE Score


The most effective solution for guaranteeing a reliable and up-to-date QE Score is often a hybrid approach, combining the two modes of data collection: synchronous for critical data that is readily available in real time, and asynchronous for other data.


The most effective solution for guaranteeing a reliable and up-to-date QE Score is often a hybrid approach

🧩 How does it work?

  • For some data, synchronous mode is simple to set up. For example, at the end of a CI/CD pipeline, an automatic call is made to send SonarQube results (technical debt, test coverage, etc.) directly to the database or Google Sheet.


  • For data that does not yet have a sending mechanism or that is more difficult to integrate, we supplement this with an asynchronous mode, based on regular scheduling.


📅 Frequencies adapted to each type of data :


  • Code security: security flaws can be recovered several times a day.


  • Functional and performance tests: daily recovery is sufficient, often after a scheduled night-time run.


  • Project quality (e.g. open tickets, bugs): every 2 or 4 hours, depending on the pace of the project.


  • GitLab activity: 1 to 2 times a day to track commits, MRs, etc.


✅ Advantages of this approach:

  • Enables a gradual transition to synchronous, without implementing everything at once.


  • Optimises the technical effort: resources are invested as a priority where the impact on data freshness is greatest.


  • Avoids wasting resources on data that doesn't evolve much.


⚠️ Points to watch out for:

  • Requires good organisation: you need to map the data sources, monitor how often they are updated, and make sure you don't forget any data.


🔐 Conclusion: The success of the QE Score depends as much on the quality of the calculation as on the freshness and reliability of the data. A well-organised hybrid approach is therefore the key to ensuring its relevance and user confidence.


Et pour vous, quel est la meilleure solution de planification de collecte de la données ?

  • hybride

  • asynchrone

  • synchrone

  • autre approche




Comments

bottom of page