Replace direct kubectl deployment with GitOps pattern:
- Build and push to git.mcintire.me registry using built-in token
- Commit updated image tag in k8s/deployment.yaml after each build
- FluxCD picks up the manifest change and handles the k8s rollout
- Remove Tailscale, kubectl, and ghcr.io dependencies
Add insecure-skip-tls-verify flag to kubectl config to bypass certificate
validation when connecting via Tailscale IP. The K3s certificates don't
include the Tailscale IP in their SANs, causing TLS verification to fail.
Previously, the deployment timestamp was set to DateTime.utc_now() when
the application started, which meant it would always show as recent in
k8s environments where pods are recreated on each deployment.
This commit:
- Updates Release.init() to check for DEPLOYED_AT environment variable
- Falls back to current time if env var is not set (backwards compatible)
- Adds proper ISO8601 timestamp parsing with error handling
- Updates GitHub Actions deploy workflow to set DEPLOYED_AT during deployment
- Creates k8s patch file for future manual deployments
The deployment timestamp will now persist across pod restarts in k8s,
showing the actual deployment time rather than pod startup time.
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Re-add Docker Buildx setup (required for cache)
- Use 'type=inline' for cache-to (embeds cache in image)
- Cache from the latest image instead of separate buildcache tag
This fixes the 'Cache export is not supported' error.
Co-Authored-By: Claude <noreply@anthropic.com>
- Remove BuildKit advanced features that were slowing builds
- Remove multi-platform builds (only build for amd64)
- Remove security scanning stage
- Remove complex caching mounts
- Simplify to basic 2-stage build
- Use registry cache instead of GitHub Actions cache
This should reduce build time from 10+ minutes to 2-3 minutes.
Co-Authored-By: Claude <noreply@anthropic.com>
- Cancels any in-progress deployment when a new commit is pushed
- Prevents multiple deployments from running simultaneously
- Uses concurrency group based on branch ref
- Saves CI/CD resources and prevents conflicting deployments
Co-Authored-By: Claude <noreply@anthropic.com>
The CI/CD pipeline now initiates the deployment but doesn't wait for it to complete, allowing the workflow to finish quickly even if pods take time to start.
Co-Authored-By: Claude <noreply@anthropic.com>
Update kubectl commands in deploy workflow to target statefulset
instead of deployment, matching the recent migration to StatefulSet.
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
Force k3s to pull the latest image by adding rollout restart
after setting the image. This ensures the cluster always runs
the most recent build even when using the :latest tag.
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Changed kubectl deployment command to use :latest tag
- Updated k8s deployment manifests to use ghcr.io/aprsme/aprs.me:latest
- This ensures k3s always pulls the most recent image from GitHub Container Registry
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
The deployment was using raw SHA but the Docker metadata action creates
tags with 'main-' prefix for the main branch. Updated kubectl command
to use the correct tag format: main-${github.sha}
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
