Docker Swarm's default FailureAction is "pause". When a task fails or is
terminated early during a rolling update, Swarm pauses the update and
stops ALL reconciliation — orphan containers persist indefinitely, even
when healthy. This is the root cause of orphan container issues reported
in production (services showing Replicas: N/1 with multiple healthy
containers that never get cleaned up).
Setting FailureAction to "rollback" makes Swarm automatically revert to
the previous working service spec on failure, preventing orphans while
preserving service availability. Also adds a default RollbackConfig with
Order: "start-first" to match the update config (Docker defaults rollback
to "stop-first" otherwise).
Only affects the default config — users who have configured their own
updateConfigSwarm/rollbackConfigSwarm are not affected.
Relates to #1669, #2223, #2911, #2150
Change Drizzle column mode from "bigint" to "number" for stopGracePeriodSwarm
across all 6 service schemas. This fixes JSON.stringify failing silently in the
@dokploy/trpc-openapi adapter, which unlike the tRPC endpoint does not use
superjson and cannot serialize BigInt values.
No database migration needed — only the JS representation changes. The values
are nanosecond grace periods that fit safely within Number.MAX_SAFE_INTEGER.
Also adds onError logging and export const config to the OpenAPI route handler
to match the tRPC route and improve debuggability.
Fixes#3793
- Updated the createNetwork function to ensure serverId is required when in cloud mode, improving error handling for network creation.
- Added braces for clarity in the conditional statement, enhancing code readability.
- Introduced a new column "networkIds" of type text[] with a default value of an empty array to the application, compose, mariadb, mongo, mysql, postgres, and redis tables.
- Updated the application and compose schemas to include the new networkIds field in their respective TypeScript definitions.
- Added corresponding entries in the migration journal and created a snapshot for version 7 to reflect these changes.
- Eliminated the unused DOKPLOY_SERVER_VALUE constant from the network handling component.
- Updated the default serverId to undefined in the network form values and adjusted related logic to ensure compatibility with cloud environments.
- Enhanced server validation in the createNetwork function to require a serverId when in cloud mode, improving error handling for network creation.
- Added components for handling and displaying Docker networks, including creation, editing, and listing of networks.
- Introduced a new API router for network operations, integrating with the database schema for network management.
- Updated the sidebar layout to include a link to the networks dashboard, ensuring user access to network features.
- Created necessary database migrations for the network table and its associated types.
- Enhanced the dashboard layout to support the new network management interface.
- Updated pnpm version from 9.12.0 to 10.22.0 in Dockerfiles for base, cloud, schedule, and server.
- Adjusted package.json files in multiple applications to reflect the new pnpm version.
- Updated dependencies for better-auth and related packages to version 1.4.18.
- Updated @better-auth/sso and better-auth dependencies to version 1.5.0-beta.16 in package.json files across multiple applications.
- Incremented @better-auth/utils version to 0.3.1.
- Adjusted pnpm-lock.yaml to reflect the updated dependency versions.
- Updated .nvmrc and Dockerfiles to use Node.js 24.4.0.
- Modified package.json files across multiple applications to reflect the new Node.js version and updated @types/node dependency.
- Updated better-auth and @better-auth/sso dependencies to version v1.5.0-beta.16 in relevant package.json files.
- Removed outdated .nvmrc file from apps/dokploy.
Mauricio Siu
Jaime Herrero
Hootan
Vibe Code
autofix-ci[bot]andGitHub
Jiří Vrba
Mauricio SiuandGitHub
mhbdev