iTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
In the FoxyInvoice deployment setup described by Lith SEO, a push to main reaches production only after parallel test, repository-conformance, and secret-scanning checks pass. The deployment then runs serially over SSH, rebuilds the application, and checks the resulting containers and live endpoints. The important distinction is that a green workflow is not proof that the intended build is serving: one masked build failure left stale code running for eight hours.
How a push moves through the FoxyInvoice pipeline
Lith SEO describes a solo-operated setup in which main is production and there is no staging environment. That is a deliberate tradeoff for this system, not a general recommendation: maintaining staging can itself introduce drift, while direct production releases leave less room to catch problems in a separate environment. The right choice depends on the application’s risk, team, and ability to test safely.
The pipeline follows a sequence of gates and verification steps:
Recommended Free Tools
- Push: A change to the default branch triggers the workflow.
- Run pre-deployment gates in parallel: Unit tests and integration tests run alongside repository-conformance checks and Gitleaks secret scanning.
- Deploy serially: After the gates pass, the workflow connects to the host over SSH and deploys the intended revision.
- Verify containers and application: The deployment waits for API containers to report healthy, rebuilds and swaps the single-page application (SPA), checks both
/healthzendpoints, and runs a smoke test. - Notify IndexNow: The workflow sends an IndexNow notification after deployment.
What the gates check
The unit and Testcontainers-backed integration tests include temporary Postgres instances and tenant-isolation coverage. Repository conformance blocks newly introduced violations while permitting existing baseline debt. Gitleaks scans for credentials. These checks provide different kinds of protection; none substitutes for verifying the deployed application.
#1 Best Overall
Why the host is reset to the intended revision
The deployment resets the host repository to the revision being released before rebuilding. This avoids using stale code if someone has manually run docker compose up while the host repository was mid-flight. The application is built on the host rather than pulled from an image registry.
How the build token is handled
A private feed token is passed to the image build as a BuildKit secret, rather than being baked into an image layer or its history. Build secrets help avoid leaving credentials in the resulting image, but the workflow must still stop if the build cannot use them.
What the deployment checks establish
In the described implementation, EF Core migrations run when containers start. The workflow waits for API container health, rebuilds the SPA on the host, swaps the generated output into Caddy’s directory, checks both /healthz endpoints, and performs a smoke test. These are the author’s reported implementation details, not an independent audit of the live service. GitHub’s documentation describes continuous deployment as automated publishing and deployment, typically after building and testing; it also documents workflows triggered by default-branch pushes. GitHub Docs: Continuous deployment
Do these 3 things before closing this tab:
1Repair Windows errors before they cause bigger problems2Scan for outdated or missing drivers - takes under a minute3Clear out junk files and repair common Windows errorsWhy a green workflow can still serve old code
The clearest failure in Lith SEO’s account was a build step that used || echo after a failure, followed by up -d --no-build. When a missing token caused the build to fail, the deployment continued with old containers. The workflow appeared successful while stale code remained active for eight hours.
The practical rule is strict: a failed build must fail the deployment. Do not suppress the build’s exit status and then start containers without rebuilding. After release, confirm that the application is healthy and that the expected revision is actually serving; workflow completion alone does not establish either fact.
How serialized deployment and runner failures differ
Serialize production releases
Production deployments should not interleave: two pushes deploying at once can compete over the same host and release state. Lith SEO describes serializing SSH deployments. GitHub Actions also supports deployment concurrency controls, but a concurrency lock can itself become a point of failure. GitHub documents concurrency and environments that can require approval or restrict branches; those are platform options, not claims about controls used by FoxyInvoice. GitHub’s continuous deployment documentation
Distinguish a stuck lock from a busy runner
One cancelled run in the author’s system did not release its concurrency group. Later runs showed zero jobs while waiting. The diagnostic clue was the combination of no jobs and idle runners, which pointed to a stuck lock rather than ordinary queueing. Check whether a run is waiting on concurrency and whether any runner is actually processing work before treating every delay as a capacity problem.
Check the labels a workflow requests
A quoted string was parsed as one literal runner label rather than a list of labels, so no self-hosted runner matched. The correction was to express the labels as a YAML list. When a job never starts, inspect the labels the workflow actually requests and compare them with the labels assigned to available runners.
Account for shared self-hosted runner queues
The described organization shares self-hosted runners across sibling repositories. The author says this followed exhaustion of GitHub-included minutes under a $0 spending limit, and that normal queueing is monitored. This is specific to that organization; it is not a general requirement to use self-hosted runners or a statement of GitHub-wide pricing. Self-hosting trades hosted-runner usage for responsibility for runner maintenance and availability, while shared capacity can add queue time.
Rollback is harder when database migrations have run
For an application rollback, the author describes resetting the host repository to a previous SHA and redeploying; because the host builds the application, this does not require an image registry. That code rollback does not automatically reverse database changes. Lith SEO’s operating practice is to fix forward rather than assume an applied migration can safely undo itself.
Avoid out-of-band production schema edits
The author reports that two manual production database edits led to startup crash loops when a later migration collided with them. If an emergency manual schema change cannot be avoided, the chapter calls for an idempotent follow-up migration and reconciliation of the migration history table. Treat the migration history and actual schema as a single operational concern: a code rollback does not make an untracked schema edit disappear.
Quick wins for a faster PC:
Repair Windows errors before they cause bigger problemsFix Now →Scan for outdated or missing drivers - takes under a minuteDriver Scan →Clear out junk files and repair common Windows errorsFree Scan →When direct-to-production deployment is a reasonable tradeoff
A pipeline without staging can suit a small, closely operated service if its gates are meaningful and production verification is dependable. It is less comfortable when releases carry high impact, multiple teams share release responsibility, or production behavior cannot be tested safely. A staging system can catch issues before production, but only if it remains representative and actively maintained; the author’s concern is that a drifting staging environment can become an obstacle rather than a useful safety net.
Best Value
GitHub environments can be configured with approval requirements and branch restrictions, and GitHub documents limiting secret access and using OIDC authentication with supported cloud providers. Those features are possible controls to evaluate; the FoxyInvoice account does not establish that this pipeline uses them. Choose controls that address the actual deployment boundary and credentials in your system.
Operational checklist for a push-to-deploy workflow
- Run tests, repository checks, and secret scanning before deployment.
- Make any failed build or deployment command stop the workflow; never continue with a stale image after a build error.
- Serialize production deployments and monitor waiting jobs, concurrency state, and runner activity.
- Pass build credentials as secrets rather than embedding them in image layers or history.
- Wait for container health and verify live health endpoints and a smoke test after release.
- Confirm the deployed revision, not just the workflow’s green status.
- Plan database changes so application rollback does not depend on an unsafe assumption that migrations reverse themselves.
The implementation details and incidents above are Lith SEO’s account of the FoxyInvoice system. Read the FoxyInvoice CI/CD chapter for the original walkthrough.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
The Tool Desk
Outbyte PC Repair FREERepair Windows errors before they cause bigger problemsFix Now →Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →
