A yearly Airflow Community survey should be run - initiated and overlooked by the PMC, and it should run proximate to the end of the calendar year and this description formalizes how we run it.

  • The intent of this survey is to understand the use of the software and features used in order to understand where to focus future development on and which features might be deprecated. But not limited to. We are also interested in perceived quality and biggest  problems for which we even might need to elaborate a solution space.
  • The questions for which are discussed on the dev list. We should have a final lazy consensus to the dev list (or vote if we don’t see consensus) on the questions before the survey is published. No formal vote is required before finalising the questions or starting the survey. It would be good to keep a majority of questions stable such that we can see changes year-over-year.
  • Questions must not involve collecting PII (Personally Identifiable Information) data. One exception is the optional collecting of email addresses for delivering sponsor incentives (i.e. training course, credits. See further on in this document)
  • There is no formal position or person responsible for generating the questions. As with everything ASF related, all individuals are given the opportunity to participate, but their influence is based on publicly earned merit.
  • The PMC will promote the survey via airflow.apache.org website - including the banners on the website, and Apache Airflow Social Media, Slack and similar channels.
  • The survey should be conducted in a way such that not one person or company gets more information than the others.
  • The results should be processed/analysed and a summary published on the Airflow website (and thus subject to normal PR review process by committers and the PMC).
  • The full, raw results for the survey should be published in full each year for the entire community to read and benefit from. Publish must be timely within 2 months after closure. The only scrubbing/redacting allowed is of obvious fraudulent answers. PII should be removed.
  • In case there is no appropriate survey platform run by the ASF available for under the “apache.org” URL, the entity or people running the survey will be free to host it elsewhere.
  • If any entity wants to support/sponsor the survey and take the cost connected with running, processing the survey - we will welcome such sponsorship. This needs to be an explicit request after a call for sponsors to the PMC and PMC has to approve it. We will also leave freedom for the entity running the survey in the way to attract wide audience (for example offering credits or free products as long as they do not suggest being PMC endorsed; and to refer Apache Airflow
    according to the nominative fair use. We will make a public call for participation in the survey so that anyone will have a chance to participate in running it. PMC has the final decision on choosing the entity to run it.
  • In case of such sponsorship, the entity will be listed as sponsor permanently  in the published Survey results - this is in accordance with the  targeted-sponsorship policy of the ASF. We will inform Fundraising of the ASF about this being a formal targeted sponsorship by the PMC.


  • No labels