- Removes coolify-testing-host.yml too, it published the testing host to Docker Hub for docker-compose.windows.yml alone, the development setup builds it locally
- Retry CDN requests only for transient errors and honor Retry-After
(Cloudflare error 1200 sends 60 s), with jitter.
- Read helper and Sentinel versions from the local versions.json, which
the hourly update check refreshes, instead of fetching per job/server.
- Replace the random 0-600 s delay with a stable per-instance offset
inside the cron interval (max 1 h).
- Manual update checks and upgrades make one CDN attempt only.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Check enabled volume backup schedules hourly and notify teams when no
execution occurs within the configured number of days. Expose the alert
threshold in the UI and API, with alerts disabled by default.
Route volume backup results and warnings to the appropriate Telegram
backup threads.
Default to sequential dispatch on self-hosted instances and concurrent
per-type dispatch on Coolify Cloud. Allow SCHEDULED_JOBS_DISPATCH_MODE
to override the default, falling back for invalid values.
Cover defaults, overrides, and fallback behavior in scheduler tests.
Add an hourly cleanup:database-import-uploads command that removes staged
import uploads older than a day, while holding the import lock of the
database so a running import keeps its file.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The scheduler ran one dispatcher for database backups, tasks, volume backups,
and Docker cleanups after each other. A slow type, for example many tasks due
at the same minute, made the whole run longer, and the shared overlap lock
then made all types skip the next minute.
Run `scheduled:dispatch --type=<type>` every minute for backups, tasks,
volume-backups, and docker-cleanups. Each type runs in its own background
process with its own scheduler and job overlap locks, and publishes and
recovers only the deliveries of its own job type. Running the command without
a type still processes all types.
The missed daily backups on Coolify Cloud came from Redis evicting queue keys
(allkeys-lfu on a full instance), not from the scan of all schedules. The
next_run_at design (452259aac, 5c5740533) did not prevent that, and it added
stored state that must be kept in sync on every frequency, timezone, and server
change. It was not released, so remove it:
- revert next_run_at, its model hooks, the "Next run" UI, and the diagnostics
changes; restore scheduled_job_states and the queue overlap lock
- a new migration removes the next_run_at columns and creates
scheduled_job_states again where the removed migration already ran
Keep the changes that fix real problems:
- run the manager every minute from the scheduler with `scheduled:dispatch`,
so a busy queue cannot delay it past the catch-up window
- load the server chain of scheduled backups with each chunk, which removes
about 7,000 queries per run on Coolify Cloud
- catch Throwable for each schedule, so one broken schedule does not stop the
others
- publish database and volume backups again when their queued job was not
started after 60 minutes (for example, lost in Redis); log lost tasks and
Docker cleanups as missed. A lost backup is not run again when a newer
occurrence of the same schedule exists.
Re-test S3 storages marked unusable every hour so a storage that becomes
reachable again is usable without manual action. Failures keep the
storage unusable and reuse the existing once-per-outage email.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The scheduled job manager checked every enabled schedule each minute and
evaluated its cron expression in PHP. On Coolify Cloud one run took more than
60 seconds, so runs overlapped and minutes were skipped.
Each schedule now stores its next due time:
- next_run_at on scheduled database backups, scheduled tasks, and volume
backups, and docker_cleanup_next_run_at on server settings (UTC, indexed)
- model hooks calculate it when a schedule is created or its frequency or
enabled flag changes; a server timezone change or a move to another server
calculates it again for the affected schedules
The `scheduled:dispatch` command runs the manager every minute from the
scheduler, not from a queue. It selects only due rows and claims each
occurrence with an atomic update of next_run_at, together with its delivery
record, so parallel runs on several nodes never dispatch it twice.
- database and volume backups run once when they are late; tasks and Docker
cleanups more than 10 minutes late are logged as missed
- a local time that occurs two times when the clocks go back runs once
- the migration continues each schedule after its last recorded occurrence and
drops scheduled_job_states
Show the next run, in the server timezone, on the database backup, scheduled
task, volume backup, and Docker cleanup pages, or "Invalid frequency" when the
frequency cannot be calculated. scheduled:diagnostics now lists overdue
schedules and open deliveries.
Remove workarounds that are no longer needed: the queue overlap lock and its
stale-lock repair, the catch-up window, and debug-only logs.
An hourly command releases managed DNS records whose references all point
to missing or deleted resources, or that have no references after a grace
period. Each record goes through the per-hostname lock and re-check, so a
hostname that a user adds back is kept. An empty reference list no longer
deletes all references of a record.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Add database-backed schedule states and deliveries so distributed schedulers publish each occurrence once, queue jobs claim executions atomically, and stale occurrences are cleaned up.
API log endpoints and the GetLogs UI now treat `all` as unbounded
output, with `-1` remaining as a compatibility alias. MCP still
falls back to a positive default. Traefik version checks run from
CheckForUpdatesJob instead of a weekly schedule.
Introduce V5Feature and config so V5 routes, jobs, commands, morph maps,
and model queries run only when enabled. Move V5 migrations to
migrations-v5 (loaded only when enabled), remove Flux from production
Docker/install paths, and add isolation tests.
Split V5 dashboard behavior into domain controllers and policies,
add agent token rotation/revocation, status reconciliation jobs,
ingress firewall syncing, and canvas connection APIs.
Add migrations for V5 status tracking, server capabilities, resource
connection aliases, and revoked agent tokens.
Serialize multiplexed SSH master creation per server to avoid concurrent workers spawning orphaned processes. Enable scheduled cleanup for stale mux connections and add guarded orphan process reaping with tests.
Add optional expiration to personal API tokens. Users pick a duration
(1/7/30/60/90 days or Never) at creation time. Expired tokens are
rejected by Sanctum, pruned hourly by sanctum:prune-expired, and a
team notification fires ~24h before expiry so owners can rotate
before API calls start failing.
- ApiTokens Livewire component stores expires_at from expiresInDays
- Rework issued-tokens UI from card grid to table (matches other views)
- New ApiTokenExpirationWarningJob scheduled hourly (idempotent via RateLimiter)
- New ApiTokenExpiringNotification (email/discord/telegram/slack/pushover)
- api_token_expiring added to alwaysSendEvents so users cannot silence
expiry warnings from the per-event notification toggle UI
- sanctum:prune-expired cadence moved from daily to hourly
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Remove unused server filtering logic in Kernel.php that was querying servers
but never using the results. Simplify Sentinel update checks in ServerManagerJob
by reusing the $isSentinelEnabled variable and removing unnecessary timezone
parameter for hourly cron execution.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
Add CleanupOrphanedPreviewContainersJob that runs daily to find and remove any PR preview containers that weren't properly cleaned up when their PR was closed.
The job:
- Scans all functional servers for containers with coolify.pullRequestId label
- Checks if the corresponding ApplicationPreview record exists in the database
- Removes containers where the preview record no longer exists (truly orphaned)
- Acts as a safety net for webhook failures or race conditions
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Add automated Traefik version checking job running weekly on Sundays
- Implement version detection from running containers and comparison with versions.json
- Add notifications across all channels (Email, Discord, Slack, Telegram, Pushover, Webhook) for outdated versions
- Create dismissible callout component with localStorage persistence
- Display cross-branch upgrade warnings (e.g., v3.5 -> v3.6) with changelog links
- Show patch update notifications within same branch
- Add warning icon that appears when callouts are dismissed
- Prevent duplicate notifications during proxy restart by adding restarting parameter
- Fix notification spam with transition-based logic for status changes
- Enable system email settings by default in development mode
- Track last saved/applied proxy settings to detect configuration drift
- Added a new job, ServerPatchCheckJob, to handle server patch checks and notifications.
- Introduced a new notification class, ServerPatchCheck, for sending updates via email, Discord, Slack, Pushover, and Telegram.
- Updated notification settings models to include server patch notification options for email, Discord, Slack, Pushover, and Telegram.
- Created a migration to add server patch notification fields to the respective settings tables.
- Enhanced the UI to allow users to enable/disable server patch notifications across different channels.
peaklabs-dev
Andras BacsaiandClaude Opus 5.5
Péter Gyarmati
Andras Bacsai