Deploy integration builds to production Worker#72
Conversation
|
OpenTofu plan for production Plan: 4 to add, 0 to change, 0 to destroy. Changes to Outputs.OpenTofu used the selected providers to generate the following execution
plan. Resource actions are indicated with the following symbols:
+ create
OpenTofu will perform the following actions:
# cloudflare_worker.site will be created
+ resource "cloudflare_worker" "site" {
+ account_id = (sensitive value)
+ created_on = (known after apply)
+ id = (known after apply)
+ logpush = false
+ name = "charm"
+ observability = {
+ enabled = true
+ head_sampling_rate = 1
+ logs = {
+ enabled = true
+ head_sampling_rate = 1
+ invocation_logs = true
}
}
+ references = (known after apply)
+ subdomain = {
+ enabled = true
+ previews_enabled = true
}
+ tags = []
+ tail_consumers = []
+ updated_on = (known after apply)
}
# cloudflare_worker_version.site will be created
+ resource "cloudflare_worker_version" "site" {
+ account_id = (sensitive value)
+ annotations = (known after apply)
+ assets = {
+ asset_manifest_sha256 = "47085533747b4a6c9cc32a0bf85f011ea7df0d5df710db5c60c24d1dba0ae8e7"
+ config = {
+ not_found_handling = "single-page-application"
}
+ directory = "/github/workspace/dist"
}
+ compatibility_date = "2026-03-03"
+ compatibility_flags = (known after apply)
+ created_on = (known after apply)
+ id = (known after apply)
+ limits = (known after apply)
+ main_script_base64 = (known after apply)
+ number = (known after apply)
+ source = (known after apply)
+ startup_time_ms = (known after apply)
+ usage_model = "standard"
+ worker_id = (known after apply)
}
# cloudflare_workers_custom_domain.site will be created
+ resource "cloudflare_workers_custom_domain" "site" {
+ account_id = (sensitive value)
+ hostname = "charm.cloudhub.social"
+ id = (known after apply)
+ service = "charm"
+ zone_id = (sensitive value)
+ zone_name = (known after apply)
}
# cloudflare_workers_deployment.site will be created
+ resource "cloudflare_workers_deployment" "site" {
+ account_id = (sensitive value)
+ annotations = {
+ workers_message = "Deploy integration builds to production Worker"
+ workers_triggered_by = (known after apply)
}
+ author_email = (known after apply)
+ created_on = (known after apply)
+ id = (known after apply)
+ script_name = "charm"
+ source = (known after apply)
+ strategy = "percentage"
+ versions = [
+ {
+ percentage = 100
+ version_id = (known after apply)
},
]
}
Plan: 4 to add, 0 to change, 0 to destroy.
Changes to Outputs:
!~ custom_domain = "dev.cloudhub.social" -> "charm.cloudhub.social"
Warning: Resource not found
with cloudflare_worker.site,
on main.tf line 1, in resource "cloudflare_worker" "site":
1: resource "cloudflare_worker" "site" {
The resource was not found on the server and will be removed from state.📝 Plan generated in Cloudflare Infra #207 |
Code Coverage OverviewLanguages: JavaScript JavaScript / code-coverage/vitestThe overall coverage remains at 21%, unchanged from the branch. Updated |
Deploying with
|
| Status | Preview URL | Commit | Alias | Updated (UTC) |
|---|---|---|---|---|
| ✅ Deployment successful! | https://pr-72-charm.justin-tech.workers.dev | f293bf8 | pr-72 |
Sun, 14 Jun 2026 18:58:34 GMT |
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 730599acb7
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| deploy | ||
| -c dist/wrangler.json | ||
| --preview-alias integration | ||
| --message "$PREVIEW_MESSAGE" | ||
| --keep-vars | ||
| --message "$DEPLOY_MESSAGE" |
There was a problem hiding this comment.
Now that this job performs the production deploy, keeping the old path filter means production is skipped for build inputs that are not listed there. I checked vite.config.ts and the build copies public/manifest.json, public/res, and public/locales into dist, and setup installs from pnpm-lock.yaml; pushes to integration that only change those files do not match this workflow's paths, so wrangler deploy never runs and charm.cloudhub.social keeps serving the previous build. Please include the missing build inputs or remove the path filter for this production deploy.
Useful? React with 👍 / 👎.
| deploy | ||
| -c dist/wrangler.json | ||
| --preview-alias integration | ||
| --message "$PREVIEW_MESSAGE" | ||
| --keep-vars | ||
| --message "$DEPLOY_MESSAGE" |
There was a problem hiding this comment.
For users who already loaded HTML or a service worker that references the previous hashed Vite chunks, this production deploy deletes replaced Worker assets immediately, so a lazy chunk or asset requested just after an integration deploy can 404 until the app recovers with a reload. I checked pnpm exec wrangler deploy --help, which documents --old-asset-ttl as expiring old assets after the given seconds "rather than immediate deletion"; add a nonzero TTL to this new production deploy path so fast-moving integration pushes do not strand in-flight clients.
Useful? React with 👍 / 👎.
| accountId: ${{ secrets.TF_VAR_ACCOUNT_ID }} | ||
| command: > | ||
| versions upload | ||
| deploy |
There was a problem hiding this comment.
When a tag or manual release deploy is running, this new production wrangler deploy can run at the same time because I checked .github/workflows/cloudflare-web-deploy.yml and it uses the separate cloudflare-infra concurrency group. Both workflows now promote the same charm Worker, so an integration push during a release can overwrite it (or be overwritten) depending only on which job finishes last; put both production paths in the same concurrency group or otherwise gate one while the other is active.
Useful? React with 👍 / 👎.
1 participant
Description
Deploys
integrationbuilds directly to the production Cloudflare Worker for the current fast-moving hosting setup, while keeping the tag/manual OpenTofu path documented for a future stable release cadence.This updates the integration deploy workflow from preview-version upload to
wrangler deploy, points the OpenTofu custom-domain default atcharm.cloudhub.social, enables Worker observability/persisted invocation logs, removes the deprecated custom-domainenvironmentsetting, and adds Cloudflare static asset headers for cache behavior and Sentry browser profiling.Pre-PR research: searched CloudHub-Social/Charm PRs and issues for Cloudflare, Wrangler, OpenTofu, deploy, and
charm.cloudhub.social. No related issues were found; nearby PR history was general repo/release/dependency work rather than an overlapping deployment change.Production deployment variable updated outside this PR: the GitHub
productionenvironmentCLIENT_CONFIG_OVERRIDES_JSONnow matches the CloudHub config for element call, push, session sync, sliding sync, polls, and featured community servers. Sentry DSN remains a GitHub secret consumed by the workflow.Fixes #
Type of change
Checklist:
AI disclosure:
This PR was generated by Codex. It changes the GitHub Actions integration deployment from uploading an aliased Worker preview version to deploying the built Worker to production with
wrangler deploy, updates the OpenTofu custom-domain default for the Worker tocharm.cloudhub.social, enables Cloudflare Worker observability, removes a deprecated OpenTofu provider attribute, records the provider checksum needed for local OpenTofu validation, adds Cloudflare_headersfor static asset caching and JS profiling, adds a patch changeset, and updates the infrastructure documentation to describe both the fast-moving production path and the stable tag/manual fallback.Validation:
tofu -chdir=infra/web fmt -checkTF_DATA_DIR=$(mktemp -d) tofu -chdir=infra/web init -backend=false -input=false && TF_DATA_DIR=$TF_DATA_DIR tofu -chdir=infra/web validate -no-colorpnpm typecheckpnpm buildtest -f dist/_headersjq '.observability, .assets, .name' dist/wrangler.jsonpnpm exec wrangler deploy -c dist/wrangler.json --keep-vars --message dry-run --dry-runpnpm exec oxfmt --check .github/workflows/cloudflare-dev-deploy.yml .github/workflows/cloudflare-web-preview.yml infra/README.md infra/web/main.tf vite.config.ts public/_headers .changeset/fast-moving-worker-production-deploy.mdgit diff --check