You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
[Consul] check option success_before_passing is not honored independently of the value it is configured to. As soon as the first check is successful the instance is marked as passing. Furthermore the first check delay doesn't seem to be guided by interval making it impossible to deterministically define a waiting period before marking a service as ready.
The expected result would be that the instance is marked critical for interval * success_before_passing and only after marked as passing.
Making that change to success_before_passing would be a breaking change, but we can introduce a different field to delay the first health check and prevent it from going to passing.
Use Case(s)
See linked issue and PR.
The text was updated successfully, but these errors were encountered:
@dnephin: there's a very old enhancement request which feels related: #954. The framing there is mainly about the option to decouple the timing of the first health check from the first interval. IIRC, right now, the first health check is staggered across the first internal to avoid a thundering herd. But some people in that thread had really long intervals for certain kinds of checks and wanted to have the initial check happen faster (or ~immediately).
See #10256 and hashicorp/nomad#9189 for more context
Feature Description
Copied from hashicorp/nomad#9189
Making that change to
success_before_passing
would be a breaking change, but we can introduce a different field to delay the first health check and prevent it from going to passing.Use Case(s)
See linked issue and PR.
The text was updated successfully, but these errors were encountered: