Service Environment step stage failure strategy retry
this fiscal quarter
D
Driving Heron
Stage level retry strategy does not work as expected when the built in "Service" step fails. In our scenario we are attempting to pull a "mirrored" image from artifactory, we expect the first attempt to pull the image will fail hence we need to pipeline to retry on failure however that is not happening.
Log In
Harness Engineering
this fiscal quarter
This post was marked as
in progress
F
Flax Raven
Along with the retry we expect to have Manual intervention option to be given when the pipeline fails at built int steps. In our case we are seeing Manual Intervention option only when deployment step fails but not on inbuilt steps.
Jeremy Nelson
Agree, that we should be respecting the stage or at least providing some failure-strategy ability around our built-in service steps, this feels like a limitation/gap(eg; Artifactory edge instances as Ryan mentioned will fail if the upstream image exists on the first pull even though it does exist if my understanding is correct). Voting this up.
D
Driving Heron
additionally stage level variables are not resolved prior to executing service and environment tests. For example if artifact and/or environment/infra depend on variables within the current deployment stage, said checks will fail.
D
Driving Heron
The built-in service and environment/infra steps do not respect failure strategy or conditional execution. For example if I have a stage conditionally executed based on variables determined in a previous stage the step can/will fail service and environment tests and we have no ability to force the stage to be skipped or to ignore the failure.
Shylaja Sundararajan
Hi Ryan,
Could you share additional details on what is the issue faced for Stage level retry strategy .Alternately manually failed Service Stages can also be re executed from the Pipeline execution run dialog
Regards
Shylaja
Shylaja Sundararajan
long-term