temporal with sql visibility has a hard limit of 3 text search
attributes per namespace. postiz requires more, causing startup
failure. adding elasticsearch enables advanced visibility mode
which removes this limitation.
add temporal and temporal-ui services to the postiz profile for
workflow orchestration. includes caddy reverse proxy with basic
auth, secret generation, and welcome page integration.
adds caddy-addon mechanism for custom certificates when let's encrypt
is not available. includes setup script with interactive wizard,
example configs, and documentation.
add cidr blocks and port-specific entries to gost_no_proxy to fix
redis connectivity issues when proxy is enabled. different libraries
interpret no_proxy differently, so we now include multiple formats.
also fix open-webui healthcheck to bypass proxy for localhost checks.
rebrand backup service following upstream project rename.
updates docker image to databasus/databasus:latest,
adds healthcheck, and includes cleanup function for
migration from old container name.
- start postgresql first, then run init_databases.sh, then other services
- add depends_on postgres for lightrag service
- remove unused nocodb from databases list
- document database initialization in add-new-service workflow
add init_databases.sh script that creates dedicated databases for
langfuse, lightrag, nocodb, postiz, and waha during install/update.
update connection strings to use service-specific databases instead
of sharing the default postgres database.
- change gost username from email to fixed 'gost' to avoid @ symbol
breaking proxy url parsing
- fix healthcheck to make real proxy request to httpbin.org instead of
invalid spider request to proxy port
- increase healthcheck interval to 60s and timeout to 15s
gost now always requires an external upstream proxy to function.
wizard prompts for upstream proxy url when gost is selected.
if no upstream provided, gost is removed from selection.
- add gost http/https proxy service with authentication
- create x-proxy-env anchor for shared proxy configuration
- apply proxy settings to ai services (n8n, flowise, langfuse, etc.)
- add gost to wizard, welcome page, and final report
- generate gost credentials and proxy url in secrets script
- use wget healthcheck compatible with alpine-based gost image
Open-WebUI can take a long time to initialize on first run, causing
the docker compose --wait to fail. Added explicit healthcheck with
120s start_period to allow proper initialization time.
replace terminal-based final report with web-based welcome page that
displays service credentials, hostnames, and quick start guide.
- add welcome/index.html with tailwind css and dark mode support
- add welcome/app.js with service metadata and password toggle/copy
- add scripts/generate_welcome_page.sh to generate data.json from env
- simplify 07_final_report.sh to show welcome page url and make commands
- add welcome page basic auth credentials to caddy and secret generation
- update add-new-service documentation with new welcome page steps
n8nio/runners image uses pnpm, not npm. configure allowed modules
via n8n-task-runners.json instead of environment variables.
- switch from npm to pnpm in Dockerfile.runner
- add n8n-task-runners.json for module allowlist config
- remove redundant env vars from docker-compose.yml
- update documentation to reflect new config location
n8n v2.0 executes code nodes in a separate task runner container.
the runner image (n8nio/runners) doesn't include npm packages like
cheerio, axios, moment, lodash - causing "module is disallowed" errors.
- add Dockerfile.runner with npm packages for code nodes
- rename Dockerfile to Dockerfile.n8n for clarity
- update docker-compose.yml to use custom runner build
- fix quote parsing in N8N_WORKER_COUNT from .env
- remove unused N8N_RUNNER_COUNT references
- parameterize n8n version for both n8n and runner images
- add healthchecks to n8n main and worker services
replace replicated workers/runners with dynamically generated 1:1
worker-runner pairs where each runner connects to its worker via
localhost (network_mode: service). this improves task routing by
ensuring each worker has dedicated runner capacity.
add generate_n8n_workers.sh script to create docker-compose.n8n-workers.yml
upgrade n8n from 1.123.3 to 2.0.0-rc.4 with external task runners for
Code node execution. task runners handle javascript and python code
execution in isolated containers with configurable concurrency.
key changes:
- add n8n-runner service using n8nio/runners:2.0.0-rc.4
- configure runner auth token and broker communication
- add N8N_RUNNER_COUNT for scaling runner replicas
- move code node library config to runner container
- update binary data mode from filesystem to database
- add runner count prompt to installation wizard
- Introduced N8N_LOG_LEVEL and N8N_LOG_OUTPUT variables in .env.example for logging customization.
- Updated docker-compose.yml to support dynamic assignment of logging configuration, defaulting to 'info' and 'console' respectively.
- Added EXECUTIONS_MODE variable to .env.example for clarity on execution settings.
- Updated docker-compose.yml to allow dynamic assignment of EXECUTIONS_MODE, defaulting to 'queue' if not set.
add environment variables for vlm pipeline support (remote services,
model loading, device selection) and persistent cache volume. improve
final report with detailed vlm pipeline usage example showing how to
use ollama with granite3.2-vision model for document conversion.
- Updated the healthcheck command in docker-compose.yml for improved readability by formatting it as a list.
- Standardized the output for PostgreSQL host in 07_final_report.sh to remove environment variable references, ensuring consistent messaging.
- Changed the Docling image reference in .env.example and docker-compose.yml to use the correct naming convention for the CPU version, ensuring consistency across configuration files.
