← All questions

Every API scenario fails at once with a connection error — where do I start?

apiRequestContext.get: connect ECONNREFUSED 127.0.0.1:3000

3 answers

• 31votes✓

  Accepted answer

  Do not bisect. Two signals in your paste say this is not the suite:

  • All of them failed, not some.
  • The failure is a connection error, not an assertion failure.

  Nothing is listening on that port. Either the application is not running, or baseUrl in envs/local.yaml (or whichever env you selected) points somewhere the app no longer is. Start the app, or check the env file, and run again.

  A suite that fails on assertions is telling you about the product. A suite that fails on connections is telling you about the environment, and that distinction saves an hour every time you respect it.

  HWanswered by Hsu Wei-lin15,230 · 22 Aug 2023
• 17votes

  Adding the third category, since these three cover almost everything:

  • exit 2 — configuration or usage. Unknown project, missing env file, unresolved ${VAR}, a bad flag value. The suite never started.
  • exit 1 with connection errors — the environment. What you have.
  • exit 1 with assertion failures — the product and the suite disagree, which is the only one worth reading scenario by scenario.

  The workshop suggests stopping the application on purpose once, precisely so you recognise the second shape when it turns up at 3am.

  RNanswered by Reeta Nandalmaintainer41,280 · 23 Aug 2023
• 12votes

  If this keeps happening on a merge gate rather than locally, take the environment out of the equation: tag the scenarios @har:<name> and run with --har-replay --strict. Strict aborts any request that is not in the HAR, so the run is genuinely offline and a gate stops depending on somebody's staging box being awake.

  JOanswered by Jide Okonkwo11,745 · 9 Jan 2024