Azure / Azure/data-api-builder

Postgresql database connection with Workload Identity in an AKS deployment

Open
#2,271 1 comment 0 reactions 0 assignees View on GitHub
cri enhancement
Dominant language
C#
Stars
1.5k
Forks
370
Avg merge
3d 22h
Merged PRs (30d)
9

Description

### What happened?

I am able to deploy the DAB container in an AKS cluster and mount the configfile, which is loaded.
Our preferred policy is to use a workload identity to authenticate with the Postgresql Flexible server database.

Connection string, via environment variable injected in dab-config.json. The client id gets replaced at deploy time via Helm:

`"Host=psqlf-demo....postgres.database.azure.com;Port=5432;Database=demo;SSL Mode=Require;User Id={{.Value.serviceAccount.clientId}}"`

Part of the config file:
` "$schema": "https://github.com/Azure/data-api-builder/releases/download/v1.1.7/dab.draft.schema.json",
"data-source": {
"database-type": "postgresql",
"connection-string": "@env('DATABASE_CONNECTION')"
},
"runtime": {
"host": {
"mode": "development"
}`

According to source code, an Azure identity is assumed when the connection string does not contain a password.
Startup fails with a 28P01: password authentication failed for

I might be missing the option to explicitly inform DAB to use Azure authentication instead of regular username/password authentication.

### Version

1.1.7

### What database are you using?

PostgreSQL

### What hosting model are you using?

Custom Docker host

### Which API approach are you accessing DAB through?

REST, GraphQL

### Relevant log output

```Text
info: Azure.DataApiBuilder.Core.Services.ISqlMetadataProvider[0]
[monsters] REST path: /api/monsters
fail: Azure.DataApiBuilder.Service.Startup[0]
Unable to complete runtime initialization. Refer to exception for error details.
Azure.DataApiBuilder.Service.Exceptions.DataApiBuilderException: Cannot obtain Schema for entity monsters with underlying database object source: monsters.monsters due to: 28P01: password authentication failed for user "b8c2bf96-..."
at Azure.DataApiBuilder.Core.Services.SqlMetadataProvider`3.HandleOrRecordException(Exception e) in /_/src/Core/Services/MetadataProviders/SqlMetadataProvider.cs:line 100
at Azure.DataApiBuilder.Core.Services.SqlMetadataProvider`3.PopulateObjectDefinitionForEntity(String entityName, Entity entity) in /_/src/Core/Services/MetadataProviders/SqlMetadataProvider.cs:line 1116
at Azure.DataApiBuilder.Core.Services.SqlMetadataProvider`3.PopulateObjectDefinitionForEntities() in /_/src/Core/Services/MetadataProviders/SqlMetadataProvider.cs:line 1054
at Azure.DataApiBuilder.Core.Services.SqlMetadataProvider`3.InitializeAsync() in /_/src/Core/Services/MetadataProviders/SqlMetadataProvider.cs:line 289
at Azure.DataApiBuilder.Core.Services.MetadataProviders.MetadataProviderFactory.InitializeAsync() in /_/src/Core/Services/MetadataProviders/MetadataProviderFactory.cs:line 65
at Azure.DataApiBuilder.Service.Startup.PerformOnConfigChangeAsync(IApplicationBuilder app) in /_/src/Service/Startup.cs:line 613
fail: Azure.DataApiBuilder.Service.Startup[0]
Could not initialize the engine with the runtime config file: dab-config.json
```

### Code of Conduct

- [X] I agree to follow this project's Code of Conduct

Contributor guide

Open the contributing guide

Research direction

Start with the PostgreSQL connection and identity handling referenced in src/Core/Services/MetadataProviders/SqlMetadataProvider.cs, especially the stack-trace locations around lines 100 and 1116. Reproduce the AKS setup using dab-config.json and the DATABASE_CONNECTION environment variable. Done means the configured workload identity can authenticate to PostgreSQL and runtime initialization succeeds without the 28P01 error.

Written by the indexing model from the issue text.

Assessment

Tech stack
azure, docker, kubernetes, postgresql
Domain
authentication, backend, cloud, databases
Issue type
Bug
Difficulty
4/5
Estimated time
3-5 days
Activity status
Stale
Clarity
Mostly clear
Newbie friendliness
35/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.