CHANGELOG.md

Changelog

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

[Unreleased]

[0.12.1-beta] - 2025-03-16

Changed

• Improve error messaging when parsing definitions

[0.12.0-beta] - 2025-03-16

Changed

• Builds with stable Rust

[0.11.1-beta] - 2021-08-28

Added

• Added a new exec command as a shorthand for running commands within the working directory of a service (eg. exec backend -- poetry install).

[0.11.0-beta] - 2021-08-22

Added

• Added a set of config cli commands (get, set, toggle, remove) for editing config file options.
• Added a mechanism with which environment variables from the users shell configuration can be periodically refreshed. This eliminates the need for the user to need to perform a daemon restart whenever they need to reload some shell configuration (eg. a change in a .zshrc file). This feature is experimental and is disabled by default.
• Added a view command to view values for all options.
• Daemon port can now be configured via the daemon.port option in config.toml.
• Add experimental interactive_shell option that when enabled and a shell command is given, the process will be spawned in an interactive shell based on the one the client is running on.

Changed

• The global configuration file has been switched from YAML to TOML (this is different from the module definitions file which will remain as YAML). This means that upon upgrading any existing config will be lost and will need to be migrated.
• Liveness status will now update correctly when a process exits.
• Probes of type exec will now timeout if their command takes more than 2 seconds.
• Daemon will now be dissasociated from its controlling terminal by default.

Fixed

• Fixed misalignment in non-emoji supporting terminals.
• Fixed white rendering as grey on light themed terminals.
• Stop group completion messages from being printed if not running in a terminal.

[0.10.2-beta] - 2021-07-14

Added

• Allow tasks to specify timeout individually rather than have a global one.

Changed

• Improve error messages for common errors.
• Error messages now print to stderr.

[0.10.1-beta] - 2021-07-13

Added

• Provide additional context for some parsing errors.

Fixed

• Fixed a bug that caused more detailed error messages to be supressed.

Changed

• Bump task timeout to five minutes (up from three).

[0.10.0-beta] - 2021-07-10

Added

• Allow services to specify services / tasks that should always be deployed "after" them, but should not constitute a direct dependency.
• Checks can now define a suggested_fix section with a command that can be run, should users choose to, in order to fix the issue the check tests for.

Changed

• daemon restart will start the daemon if it is not running

[0.9.5-beta] - 2021-07-07

Added

• Allow executable probe and shell definitions to also use the shell shortcut (instead of specifying a command).
• Add daemon restart shortcut for restarting the daemon

Fixed

• Supress output of stderr from checks

[0.9.4-beta] - 2021-07-04

Changed

• The always_await_readiness_probe option will now default to true instead of false.

[0.9.3-beta] - 2021-06-25

Fixed

• Do not skip deployment of any tasks when --force is active.

[0.9.2-beta] - 2021-06-14

Added

• Log files for tasks can now also be viewed with logs <task-name>

[0.9.1-beta] - 2021-06-14

Fixed

• Fixed an issue where where a group's tasks would always redeploy regardless of whether the services were deployed.

[0.9.0-beta] - 2021-06-12

Added

• Tasks will now no longer deploy if the service is in a healthy state and running.
• Allow check definitions to also use the shell shortcut (instead of specifying a command).
• The ps command is now coloured by default making failed services more easily stand out. This may be disabled using -n/--no-color.

Fixed

• Fixed usage of shell on tasks when ran with the run <task> option.
• Fixed a bug where the same check would be run twice when include by two different services.

[0.8.1-beta] - 2021-05-05

Added

• Services can now define environment_sets to allow configuring environment variables for a deployment. Using -e .../--env ... as extra arguments to deploy can activate one or more environment sets.
• Added a new shell option for services and tasks (on the same level as command) to make invoking shell commands less verbose.

[0.8.0-beta] - 2021-04-24

Added

• Added support for parallel deployment. The deploy command will now deploy modules in parallel by default.
• Added --threads/-t to control how many threads to use to deploy in parallel. Also, added --serial/-k to force serial deployment.

Fixed

• Fixed an issue where dependencies would be deployed in the wrong order.

[0.7.4-beta] - 2021-04-10

Added

• Added a new ordered_dependencies module option for specifying dependencies of a services that need to be run sequentially.

[0.7.3-beta] - 2021-03-30

Fixed

• Fixed a panic in the daemon that would cause readiness probes to always fail.

[0.7.2-beta] - 2021-03-28

Fixed

• Fixed an issue where the daemon would be relaunched unnessecarily.

[0.7.1-beta] - 2021-03-28

Added

• Added daemon launcher script for launching the daemon while opening shell terminals.

[0.7.0-beta] - 2021-03-27

Added

• Reworked the healthchecks system into readiness and liveness probes which can be independently used to check whether a service has started (readiness) and whether the services continues to live (liveness). Previous configuration entry of healthcheck has been renamed to readiness_probe, and a new module configuration block named liveness_probe has been introduced.
• Added a new option to ps to show liveness status.
• Allow specifying shell type to invoke using -t/--type, and adding a type on a shell definition.
• Improved error messages when specifying dependencies that don't exist.
• Improvements to the way the daemon is run to better support shell based environments.

Changed

• A service with a failing liveness probe will now always be redeployed.

[0.6.0-beta] - 2021-02-18

Added

• Added a new healthcheck of type net which allows monitoring the health of a service by trying to obtain a TCP connection. The healthcheck attempts to establish a connection, which upon established will make the healthcheck succeed. On a TCP reset the healthcheck is considered failed and will also time out after 100ms.
• Added a new -o option to allow deploying only selected modules and not their dependencies.
• Added a new -w/--wait option for enforcing all deployed modules to pass their healthcheck before continuing.
• Added a new shell command and module type for defining and opening a shell to a service.

Changed

• Renamed the module definition file extension from .yaml to .yml.

Fixed

• Fixed an issue in working directory resolution for shell definitions.

[0.5.0-beta] - 2021-01-24

Changed

• Changed flag for skipping checks to -z and --no-checks.
• Improved the error message when a healthcheck failed due to misconfiguration.
• Daemon default port changed to 13754 to reduce chance of conflict.
• Changed release status to beta.

Fixed

• Attempting to stop or restart a module that doesn't exist now returns an appropriate error message.
• Skip checks flag (-z) is now under the correct command.

Added

• Allow forcing deployments to always redeploy using -f or --force.
• Introduced a new type of healthcheck (log_line) which allows for considering a service healthy when a certain line has appeared in its stdout.
• Introduced a new client configuration file to contain persisted options.
• Add daemon_port option to allow configuring the daemon port via the client configuration file
• Add default_dir option to allow configuring the default directory to look for module definition files in. Note that this directory is always last in precedence.
• Allow skipping healthchecks using -s / --no-healthchecks.

[0.4.2-alpha] - 2021-01-17

Changed

• Minor improvements to wait spinners.

[0.4.1-alpha] - 2021-01-17

Added

• Module file can now be discovered on any path ancestral to the path the command was called.
• Modules can now be overriden with user specific settings with an overrides file (.override.yaml') or using -o <override_file>.

Fixed

• Fixed an issue where healthcheck would always run regardless of dependencies.

[0.4.0-alpha] - 2021-01-16

Fixed

• Print help when no subcommand specified.

Changed

• Healthchecks are not awaited by default. Only when there is a dependency on the module (or a post_up task is defined) then the healtheck will be awaited. This behaviour may be overriden with always_wait_healthcheck in the module definition.

Added

• Introduced two new options post and post_up to add tasks to be performed after a service has been deployed (post) / or has passed its healthcheck (post_up) respectively (experimental).

[0.3.3-alpha] - 2021-01-13

Fixed

• Fixed an issue where the paths failed to resolve if working_dir was not specified.

Added

• Made retries for a healthcheck configurable via the retries property.

[0.3.2-alpha] - 2021-01-12

Added

• Added aliases for most commands (eg. deploy,d, restart,rr, run,r, stop,s).
• Added two new options for logs --all and --follow.
• The path to the module definition file can now be specified with -f or --file.

[0.3.1-alpha] - 2021-01-11

Added

• Added the ability for the stop command to stop multiple services at once.
• Added a new down command to allow stopping every running service without having to specify names.

Changed

• Healthcheck kind has been renamed to type and it's values are now lowercase (eg. type: exec).
• Increased the number of attempts for a healthcheck by 4 seconds.

Fixed

• Paths within module definitions are now correctly resolved relative to the client rather than the daemon.
• Fixed an issue where the healthchecks command stdout will be printed on the daemon stdout.

[0.3.0-alpha] - 2021-01-11

Added

• Added a new Group declaration to allow for modules whose sole purpose is to group other modules. They can be deployed, specified as a dependency, or have checks just like other modules.
• Added cleanup for child processes on SIGINT or SIGTERM
• Services can now define a healthcheck section to implement checks which must pass before further dependencies are started (experimental).

[0.2.2-alpha] - 2021-01-07

Changed

• Changed the default pager from less to tail with a default 30 line output.

Fixed

• Fix incorrectly labeled log files. Services now have a .service suffix and tasks have .task.

Added

• Allow termination signal to be specified on services using termination_signal on the YAML config.

[0.2.1-alpha] - 2021-01-06

Fixed

• Switched to SIGKILL as the default kill signal. Fixes a race condition with contention on the log file during restarts.

[0.2.0-alpha] - 2021-01-05

Changed

• Increased the timeout for deploying tasks.
• Improved error message when trying to view logs of non existant service.

Fixed

• The daemon will now attempt to also kill descendant processes (rather than only terminating the parent and leaving orphans behind). This is only implemented for Unix based systems and is not foolproof.

[0.1.2-alpha] - 2021-01-04

Fixed

• Fixed a bug that would cause a panic during dependency resolution.

Changed

• The environment and dependencies options are now optional.
• The ps command now has better output with long service names.
• The error messages when a service fails to start are now slightly more helpful.

Added

• Resolution for the tilde in working_dir paths.
• Running with no command now prints the cli help.
• Added a changelog!

[0.1.1-alpha] - 2021-01-03

Added

• Added restart command line option for restarting services that may have failed to start or are stuck.

[0.1.0-alpha] - 2021-01-03

Added

• Initial version of the project