True Causes
TR
Bahasa IndonesiaDanskDeutschEnglishEspañolFrançaisItalianoNederlandsNorskPolskiPortuguêsSrpski (latinica)SvenskaTiếng ViệtTürkçeΕλληνικάРусскийСрпски (ћирилица)Українська中文日本語한국어
Why does the same outage keep recurring?
Rollbacks take longer than the outage they are meant to end
a walkthrough with stand-ins — a demonstration, not a finding · root cause · sealed
It significantly influences
Reliability has no budget line of its own
certified · 12 yes · 1 no · 1 abstained
What significantly influences it
Feature flags are never cleaned up, so nobody knows which paths are live
certified · 13 yes · 0 no · 1 abstained
Judged not to be linked
The same three services cause most of the pages
judged not to · 1 yes · 13 no
Config lives in five places and drifts between them
judged not to · 3 yes · 11 no
Customers report outages before monitoring does
judged not to · 0 yes · 14 no
The incident channel fills with people asking for status instead of giving it
judged not to · 1 yes · 11 no · 2 abstained
Retries are unbounded, so a slow dependency becomes a flood
judged not to · 0 yes · 14 no
The staging environment does not resemble production
judged not to · 2 yes · 11 no · 1 abstained
Feature flags are never cleaned up, so nobody knows which paths are live
judged not to · 3 yes · 9 no · 2 abstained
The service map in the wiki is two years old
judged not to · 2 yes · 12 no
The same three services cause most of the pages
judged not to · 1 yes · 13 no
Config lives in five places and drifts between them
judged not to · 6 yes · 22 no · 1 abstained · cross-group
Reliability has no budget line of its own
judged not to · 1 yes · 13 no
Customers report outages before monitoring does
judged not to · 2 yes · 11 no · 1 abstained
The incident channel fills with people asking for status instead of giving it
judged not to · 0 yes · 13 no · 1 abstained
Retries are unbounded, so a slow dependency becomes a flood
judged not to · 1 yes · 13 no
The service map in the wiki is two years old
judged not to · 2 yes · 11 no · 1 abstained
A group’s judgement, not a measurement. Demo room. On the atlas.
↗ cite or embed this claim
Link: https://truecauses.org/c/6708d57/0c1609a-rollbacks-outage-meant
As data: https://truecauses.org/c/6708d57/0c1609a.json