omctest-service-viewer - what's actually wiredSnapshot 2026-06-23, post June (#276) deploy. Zone test.dukc.net (Z1T89WN8T3VS8O, in SharedServices). 18 service A-records cross-checked against the 13 ALBs that currently exist in Test.
EAI_NODATA) is one symptom of a wider problem: 5 of 18 service hostnames point at load balancers that no longer exist. Records are updated piecemeal each cycle and after ALB churn - no single owner keeps them in sync. The recent service-viewer ALB delete+recreate left some records repointed and others stranded.
omctest-service-viewer (2020) owns the shared public ALB + 443/80 listeners that AIS/DDS/MSDUKC hang their listener-certs on. It's a full app stack (ASG + KMS key w/ default Delete policy + IAM + the ALB) - far more than "an ingress".omcte-publi-av10jm2ywvl0) as "unowned" - it wasn't; that broke every component deploy. It was recreated as omctest-svcviewer-public (new name) via a targeted change set (KMS left intact).ais/dds/msdukc* were repointed to the new ALB, but dds-ecs & serviceviewer were left pointing at the deleted one.tf-lb-..., omctes-...) and repoints some records - leaving prior-cycle records dangling (e.g. api still points at a March 2025 ALB).| Hostname | Points at (alias) | Status | Impact |
|---|---|---|---|
services.test.dukc.net | internal-test-inter-13p0cqhh92h5g (internal, deleted) | DANGLING | The calc failure. RAO endpoint services.test.dukc.net/raoservice/ -> no address -> every passage-plan calc fails. |
serviceviewer.test.dukc.net | omcte-publi-av10jm2ywvl0 (deleted in the audit) | DANGLING | Should be the recreated omctest-svcviewer-public (like ais/dds were). |
dds-ecs.test.dukc.net | omcte-publi-av10jm2ywvl0 (deleted in the audit) | DANGLING | Same orphaned-by-the-rename case. |
auth.test.dukc.net | omctes-publi-hofpv5kk5lzo (not in current ALBs) | DANGLING | Cognito custom-domain target gone. Login still works via the app form, so confirm whether it's actually used. |
api.test.dukc.net | tf-lb-2025030501... (a March 2025 ALB) | STALE ~15mo | Points at an ALB from over a year ago - almost certainly dead. |
By ALB creation time, these split in two. 7 were stood up by this cycle's deploy (#276, ALBs created 15:36-15:42 on 22 Jun). The other 6 hang off omctest-svcviewer-public - the service-viewer ALB recreated by hand at 06:37 after the audit deleted it. The deploy did not create that ALB; those 6 depend on persistent, manually-rebuilt infra that sits outside the cycle.
| Hostname | Points at | Origin |
|---|---|---|
test.dukc.net (apex/front-end) | omctes-Publi-ItOji1crmWOi | #276 cycle |
apiservices | tf-lb-...365347...0b | #276 cycle |
dukcconfigbasedservices | tf-lb-...424550... | #276 cycle |
api.service | tf-lb-...424521... | #276 cycle |
api.series6 | omctes-Publi-ItOji1crmWOi | #276 cycle |
ice | omctes-Publi-Ag5HHzaOEamg | #276 cycle |
phdr | omctes-NLBPu-dpX60G7jTeVE (NLB) | #276 cycle |
ais / dds / msdukc1 / msdukc2 / msdukc-dds2 / msdukc-melbourne1 | omctest-svcviewer-public | hand-rebuilt 06:37 - NOT the cycle |
raoservice? (6 new internal ALBs, all generic tf-lb-.../omctes-Inter-... names, custom-header routing - can't tell from AWS alone. This is the one bit of knowledge needed to fix the calc now.)