2 years ago
I recently had my CI begin to fail as a result of a failing service instance redeploy. As a quick tldr I've been using this github action https://github.com/Faolain/railway-pr-deploy/blob/main/index.js for many months now unchanged. However I recently added a new service and it appears after adding this new service, the serviceInstanceRedeploys fail via the CI (which uses the graphql endpoint). I tried on my local graphql GUI and see the same issue. Is there any reason why this serviceInstanceRedeploy is failing/how to find more information about this?
26 Replies
Something to note is that this particular service which is encountering an issue is based on a template but I'm not sure if that is the reason for the failure
Unsure why I wouldn't be able to redeploy the service like any other. There doesn't seem to be much information either about the error other than
{
"errors": [
{
"message": "Problem processing request"
}
],
"data": null
}2 years ago
mind sending the query and variables over?
mutation serviceInstanceRedeploy($environmentId: String!, $serviceId: String!) {
serviceInstanceRedeploy(environmentId: $environmentId, serviceId: $serviceId)
}{
"environmentId": "3ec1dd8f-4c44-415b-988b-955288e47bec",
"serviceId": "e63f96dd-c01f-45a5-ba40-a9f024aaa3bb"
}the other 2 services are redeploying just fine, it's just this service which is the problematic one
(it's programmatic so no change in the syntax just a loop which passes the different serviceIds through, the one based on the template is the one that's failing)
2 years ago
what is the state of the service?
2 years ago
that is not a valid state though, im talking about in the context of railway
The Github Action logs
Environment variables updated for all services.
Waiting 15 seconds for deployment to initialize and become available
Deployment is still in progress. Status: BUILDING . Waiting 20 seconds and trying again...
Deployment is still in progress. Status: DEPLOYING . Waiting 20 seconds and trying again...
Deployment is still in progress. Status: DEPLOYING . Waiting 20 seconds and trying again...
Updating Deploying Trigger to new Branch Name
Updating Deploying Trigger to new Branch Name
Updating Deploying Trigger to new Branch Name
All deployment triggers updated successfully.
Redeploying Service...
Environment ID: 3ec1dd8f-4c44-415b-988b-955288e47bec
Service ID: e3273da5-c840-4faf-8a15-8bb12c0e43d5
Redeploying Service...
Environment ID: 3ec1dd8f-4c44-415b-988b-955288e47bec
Service ID: e63f96dd-c01f-45a5-ba40-a9f024aaa3bb
Redeploying Service...
Environment ID: 3ec1dd8f-4c44-415b-988b-955288e47bec
Service ID: 4fd3601c-fd90-4195-997d-051be3ff145f
Error: Action failed with error: Error: Problem processing request: {"response":{"errors":[{"message":"Problem processing request"}],"data":null,"status":200,"headers":{}},"request":{"query":"\n mutation serviceInstanceRedeploy($environmentId: String!, $serviceId: String!) {\n serviceInstanceRedeploy(environmentId: $environmentId, serviceId: $serviceId)\n }\n ","variables":{"environmentId":"3ec1dd8f-4c44-415b-988b-955288e47bec","serviceId":"e63f96dd-c01f-45a5-ba40-a9f024aaa3bb"}}}The state of the other services are the same as the one that's failing
2 years ago
I think you should take this as a sign to update the action to use staged changes
I believe at the time when I was looking at this it was not feasible due to a lack of support of certain features that we needed
hence the railway team undeprecating one of the endpoints to maintain continued support
in any case that seems a bit over the top for what seems to be a serviceInstanceRedeploy which we may have to do later programatically in any case
2 years ago
it was possible i did it in the dashboard just fine, and i think jr even recommended you used staged changes too
^ , is there any way railway can investigate why this serviceInstance is not deploying whereas the others are deploying just fine?
2 years ago
closing in favor of your private thread!