Phantom Orphaned Services that do not appear on GUI
I had created a service on the GUI and then deleted it(I believe done twice), however now whenever I create environments based off that previous one (even though that previous one doesn't have these dangling empty services) they appear on the programatically created ones. Stranger, these services do not show on the GUI. How should I remedy this? Should I delete the serviceinstances via the graphql api? Are these dangling services consuming any resources? Etc
16 Replies
Project ID:
7b1d4989-a0d3-4bb1-98de-a5372becb82d
7b1d4989-a0d3-4bb1-98de-a5372becb82d
Test
this displays when making a graphql query from the API but the frontend only shows the one service
4fd3601c-fd90-4195-997d-051be3ff145f
with those other 2 being ghosts/phantomsi wonder why the dashboard is deciding now to show you those two services, but you can certainly try to delete them manually with the api
So to clarify the dashboard doesn't show them (i had created them in the past and deleted them) but I realized It was Broken when my GitHub action began to fail. (Tldr my GitHub action reads ".domain" from the service for preview deploys. But in this case It was failing and I wasn't sure why. It was only after investigating why .domain was null ....and that It was because there was a Phantom service instance that existed but doesn't display on the GUI
It only shows when making a graphql request
interesting, have you tried deleting them with the api?
I did not do so in case anyone from the team wanted me to keep them live to investigate how they could have gotten to that state despite not appearing in the UI
gonna have to delete them as it's blocking some other work I'm doing unfortunately but I'll post back here if/when it happens again Edit: Actually nevermind how do I delete these with the API? I tried a service delete but the serviceinstances remain. Maybe this is why when I deleted them off the GUI they remain in this stranded state bc there was some failure that's unaccounted for?
I don't see one for delete for the serviceInstances
Here are some of the stranded serviceInstances that need to be deleted
its the
serviceDelete
mutationyep I've done that but nothing happens
but when I run the get again, it remains there
will flag the team but they may only be able to get to this after the new years.
please send those two service ids in plaintext.
serviceId seems to be tied to serviceInstance, it may be possible because of an existing serviceInstance that the service is unable to delete or because it was forked from a staging environment which had it, but that no longer has it....though I'm not sure why that would be an issue unless there's a dangling connection there
yep will do!
"environmentId": "0a67cc08-c99e-4e26-9578-22c3423320d0",
"serviceId": "253284f3-8521-4fae-a151-c7ff27c7432c"
"serviceId": "a0dbd455-5ee5-4127-80a7-ecff0b73d9b4",
Thread has been flagged to Railway team by @Brody.
Hmm, this is odd. @fp can you take a look when you're back?
For context, I am back on Monday so will look at this then!