bug: Platform container incompatible with alternative images (e.g., flowai)

[wcollins]

Description

The Platform container configuration has hardcoded assumptions that prevent compatibility with alternative Platform images like flowai:v0.0.2. Three issues were identified:

1. Health check uses wget - The flowai image includes curl but not wget, causing Docker to mark the container as "unhealthy" even though the platform is running correctly.

2. Hardcoded UID 1001 - The platform-init container sets volume ownership to 1001:1001, but flowai runs as UID 1000. This causes "Log directory is not writable" errors.

3. Gateway Manager API requires gateway_name - The flowai API requires a gateway_name field when creating gateway clusters, which the standard Itential API does not require.

Reproduction Steps

1. Set PLATFORM_IMAGE=flowai:v0.0.2 in .env
2. Run make setup
3. Observe:
   • Container marked "unhealthy" (health check fails)
   • Platform logs show "Log directory /var/log/itential is not a writable location"
   • Gateway Manager cluster creation fails with "Missing gateway_name in parameter"

Expected Behavior

The dev stack should be configurable to work with alternative Platform images that have different:

• Available CLI tools (curl vs wget)
• Container UIDs
• API requirements

Environment

• OS: Linux
• Version: Docker Compose with profiles
• Platform Image: flowai:v0.0.2 (UID 1000, has curl, different API)

Solution

1. Change health check from wget to curl (more universally available)
2. Make platform-init UID/GID configurable via PLATFORM_UID and PLATFORM_GID env vars (defaults to 1001 for standard images)
3. Add gateway_name field to Gateway Manager API call (required by some Platform versions)

Checklist

• I have checked existing issues to avoid duplicates
• I have provided a clear description of the issue
• I have included steps to reproduce (if applicable)
• I have included relevant environment details