v0.16.1

What's Changed

• Update docs. by @janbjorge in #199
• Add test_execute_after_updated_gt_execute_after by @janbjorge in #200
• Task manager check task cancelled upon completed by @janbjorge in #203
• Dynamic import add current working directory to sys-path by @janbjorge in #204

Full Changelog: v0.16.0...v0.16.1

v0.16.0

PGQueuer v0.16.0 Release Notes

Release of v0.16.0, bringing new capability: deferred job execution for more precise control over job timing.

Breaking Changes / Migration Notes

• Database Schema Update: The Job model now includes a new execute_after column. This change requires a database schema update. To apply the necessary migration, please run:

python -m pgqueuer upgrade

Failing to apply this migration will result in errors when trying to enqueue jobs with the new execute_after attribute.

What's Changed

• Add dataclass dec. to job executor by @janbjorge in #194
• Refactor verify structure by @janbjorge in #196
• Add execute after feature by @janbjorge in #195.

Example Usage

To enqueue a job for deferred execution until 1 minute from now:

from datetime import timedelta
await Queries(apgdriver).enqueue("my_task", payload=None, priority=0, execute_after=timedelta(minutes=1))

This job will only be picked for execution once the specified minute has elapsed, giving you better control over job timing.

Full Changelog: v0.15.0...v0.16.0

v0.15.0

PGQueuer v0.15.0 Release Notes

Im excited to introduce a new features and enhancements in this release; support for scheduling recurring jobs using cron-like syntax.

What's Changed

• Public Interfaces Defined: Updated __init__.py to define public interfaces more clearly (#186).
• Scheduling with Crontab Syntax: You can now schedule functions using cron-like expressions for periodic task execution (#188).
• Added SchedulerManager to Public API: Included SchedulerManager in __init__.py for easy access (#191).
• Added Scheduler Example to README: Enhanced documentation with examples on using the new scheduling feature (#190).
• Batch RPS Events: Implemented batching to improve the handling of rate-per-second (RPS) events (#192).

Note: In order to use the new scheduling feature, a migration step is required. Please run:

python -m pgqueuer upgrade

Example Usage

To schedule a recurring task every minute:

@scheduler.schedule("sync_data", "* * * * *")
async def sync_data(schedule: Schedule) -> None:
    print("Running scheduled sync_data task")

Run the scheduler using the CLI:

pgq run myapp.create_scheduler

Full Changelog: v0.14.1...v0.15.0

v0.14.1

What's Changed

• Minor readme update by @janbjorge in #179
• Add tooling concurrency-probe by @janbjorge in #180
• Switch to utc_now() for consistent timing across multiple processes by @janbjorge in #182
• CLI listen broken by @janbjorge in #183
• Use buffer for RPS by @janbjorge in #181
• Refactor listener setup by @janbjorge in #184

Full Changelog: v0.14.0...v0.14.1

v0.14.0

Breaking Change

This release introduces a new database schema. Before deploying this version, ensure your database schema is updated by running the following command:

python -m pgqueuer upgrade

This step is essential for compatibility with new features, including schema modifications related to job management.

What's Changed

• PsycopgDriver Fix: Corrected missing inheritance for PsycopgDriver, enhancing type compatibility and functionality. #169
• Renamed alive to shutdown: Updated the alive attribute to shutdown to better reflect its purpose across the codebase. Backwards compatibility maintained with a deprecation warning. #168
• Exponential Backoff with Jitter: Added exponential backoff with jitter for deadlock mitigation, improving job processing resilience in high-concurrency environments. #171
• Queue Manager ID Column: Introduced a queue_manager_id column to support isolated queue management and improved tracking across multiple managers. #175
• alive Property Deprecation: Issued a deprecation warning on the alive property to encourage migration to shutdown. #176
• Async Generator for Job Dequeuing: Refactored job dequeuing logic into an async generator, enhancing efficiency and readability. #174

Full Changelog: v0.13.0...v0.14.0

v0.13.0

What's Changed

• fix: dashboard show Created in user timezone by @trim21 in #160
• Add doc-string retry_timer_buffer_timeout by @janbjorge in #161
• Add pgq cli script by @janbjorge in #164
• Update cli docs by @janbjorge in #165
• Updated readme by @janbjorge in #166

New Contributors

• @trim21 made their first contribution in #160

Full Changelog: v0.12.0...v0.13.0

v0.12.0

What's Changed

• Add pgq as cli option by @janbjorge in #159

Full Changelog: v0.11.0...v0.12.0

v0.11.0

What's Changed

• Use uv in github actions by @janbjorge in #147
• Number of consumers in test_serialized_dispatch_mixed must be gt. 1 by @janbjorge in #148
• Add sleep ci job by @janbjorge in #150
• Add executor to entrypoint fn by @janbjorge in #155
• Cleanup a few code smells by @janbjorge in #158

Full Changelog: v0.10.0...v0.11.0

v0.10.0

Im excited to announce Serialized Dispatch for PGQueuer!

Serialized Dispatch ensures jobs of the same type are processed one at a time, perfect for avoiding race conditions when modifying shared resources.

Example

Define an entrypoint with Serialized Dispatch:

@qm.entrypoint("data_processing", serialized_dispatch=True)
async def data_processing(job):
    # Job processing logic
    ...

This ensures only one data_processing job runs at a time, avoiding concurrent modifications.

Other changes

• Add validate benchmark ci test by @janbjorge in #137
• Update ci.yml by @janbjorge in #138
• Delete tools/compare_rps_main.py by @janbjorge in #139
• Add ci schedule by @janbjorge in #140
• Add serialized dispatch by @janbjorge in #136

Full Changelog: v0.9.2...v0.10.0

v0.9.2

What's Changed

• Add heartbeat docs. by @janbjorge in #134

Full Changelog: v0.9.1...v0.9.2