I just had an issue where when I run terraform apply, if I have some other resources that take a while to have their states refreshed, the kubectl_manifest ones will error out with the following message:
| Error: failed to create kubernetes rest client for read of resource: Get "https://REDACTED/api?timeout=32s": context deadline exceeded (Client.Timeout exceeded while awaiting headers)
│
│ with kubectl_manifest.redacted["ServiceAccount"],
│ on redacted.tf line 2, in resource "kubectl_manifest" "redacted":
│ 2: resource "kubectl_manifest" "redacted" {
│
I managed to go around this issue by having the kubectl_manifest resources depend (with depends_on) on some non-related resource that has its state refreshed late in the terraform plan phase.
Would it be possible instead to specify a timeout longer than 32 seconds (as can be seen in the error message above)?
I just had an issue where when I run
terraform apply
, if I have some other resources that take a while to have their states refreshed, thekubectl_manifest
ones will error out with the following message:I managed to go around this issue by having the
kubectl_manifest
resources depend (withdepends_on
) on some non-related resource that has its state refreshed late in the terraform plan phase.Would it be possible instead to specify a timeout longer than 32 seconds (as can be seen in the error message above)?