* add checkpoint restore events
* fix retries.enabledInDev
* fix tsconfig paths
* fix provider build and dev
* update kubernetes provider and fix builds again
* upgrade prod base to node 20
* update infra publish workflow
* rethrow k8s errors after logging
* shorten index container names
* fix error type assertion
* rename type assertion
* remove resource limits for now
* add missing run id on create
* add push to deploy command for self-hosting
* checkpointing fixes
* update coordinator image
* ensure valid registry login
* delete checkpoint archive after successful push
* log options on error
* structured logs for socket connections
* fix structured log merge
* exit process after checkpointing
* update restore pull secret name
* append shortcode to restore names
* log handler payload
* disable post start lifecycle hook
* pass in coordinator host via volume
* replace dapi with taskinfo
* add missing restore label
* don't restart restored containers
* remove init container from create
* atomic post-completion checkpoints
* switch to run id for container names
* improve wait accuracy
* measure basic checkpoint perf
* always log disconnect reason
* use system clock to end wait spans
* checkpoint readiness and cancel signals
* restore from checkpoint events and fix statuses
* remove attempt id env var
* restore dependencies from events
* reconnect wip
* lifecycle hooks are back
* fix hooks and improve reconnect
* make docker send postStart hook
* only checkpoint for retry if large delay
* fix a few more resume issues
* lifecycle hook fixes
* skip connection handler when waiting for post start hook
---------
Co-authored-by: Eric Allam <eric@trigger.dev>
* fix basic resume mechanics
* fix dependency resume
* fix deploy page sort
* remove task list console log
* checkpoint restore between attempts
* add checkpoint requirement checks to docker provider
* switch to raw execa for long statements
* catch and report pre-connection index errors
* graceful worker shutdown after completion
* fix final completion
* use same task process for each attempt
* update retry catalog entry
* add immediate return task
---------
Co-authored-by: Eric Allam <eric@trigger.dev>
* zod ipc connection with acks
* pass sender through to ipc handlers
* bits and bobs
* add host networking back in
* disable verbose logs
* restore after tasks and batches.. almost
* restore and resume all the things
* fix for systems without checkpoint support
* Get deployment indexing errors to work with the new ZodIpc stuff
---------
Co-authored-by: Eric Allam <eallam@icloud.com>
* WIP proxy/deploy
* WIP, registry proxy in express working
* A couple of notes, preparing for indexing
* Move the changes to prod-worker into the new file
* Deploy command working with indexing and runs (docker provider only for now)
* Removed ts-expect-error directive
* Fixed build command
* add zod socket
* start using zod socket
* use zod socket for shared queue
* use shared provider shell
* update k8s provider task ops
* set otlp endpoint
* update docker actions and fix builds
* remove unused types
* update injected secret key env var name
* complete socket.io types to schemas migration
* update env example to new v3 key var
* fix coordinator build
* set task run label
* ensure attempt id is always set
* increase prod worker timeout
* set otlp endpoint for on runs
* pass otlp env var to prod worker
* auto-remove completed index jobs
nicktrn