Released on February 5, 2021.
- Install curl in server docker image for healthchecks - #182
- Improve
LoopService
interval handling - #43
- Move to
YYYY.MM.DD
versioning scheme away fromYYYY-MM-DD
- #178
Released on January 25, 2021.
- Add 'description' field for Flow Groups along with routes for updating - #177
Released on January 19, 2021.
- Add
sum
aggregate to ORM - #174
Released on January 5, 2021.
- Support per-flow-run and per-flow-group
run_config
overrides - #166
- Upgrade Apollo NodeJS to
14.15.1
- #153 - Helm: configure strategy for deployment - #165
- Allow for timezone specification on flow group schedules - #169
- Add
flow_group.run_config
andflow_run.run_config
- #166
Released on December 16, 2020.
- Add the
events
RBAC permission to agents created by the helm chart - #157
Released on December 14, 2020.
- Add documentation on how to connect the CLI to a remote deployed server - #159
- Refactor the Zombie Killer to rely entirely on flow run heartbeats - #158
- Add support for custom Hasura root fields - #161
- Add jobs/status to role for agent - #154
Released on December 4, 2020.
- Add flag for shutting down loop services - #144
- Support configurable primary keys in ORM - #145
- Add support for expanded Hasura update operations - #146
- Helm chart: Always pull prefecthq docker images by default - #151
- Listen on all agent interfaces for healthchecks - #148
- Add task run name index for text search - #149
Released on November 29, 2020.
- Add KubernetesAgent to the Helm chart - #136
- Add two new GraphQL routes for Core functionality - #143
- Fix handling for managed postgres usernames with
@
signs - #139
Released on November 13, 2020.
- Add the Prefect Core version to the api query - #126
- Remove global 24 hour expiration on flow run idempotency keys - #134
- prefect-server dev infrastructure would sometimes fail to start with 'connection closed' - #130
- Fix helm chart when using existing postgres instance - #132
- Fix scheduling duplicate runs due to expired idempotency keys - #134
Released on November 11, 2020.
- Add experimental Helm chart for deploying in K8s - #123
Released on November 10, 2020.
- Add API for persisting and retrieving task run artifacts - #121
Released on October 29, 2020.
- Add idempotency keys to
flows.create_flow
- #116
- Expose database upgrade errors during infrastructure start - #117
Released on October 27, 2020.
- Check for existence of agent before updating flow run agent - #114
Released on October 22, 2020.
- Update flow settings graphql logic to use supported API routes - #113
Released on October 14, 2020.
- Allow for scheduling different parameters and different run labels at the exact same time - #111
Released on October 13, 2020.
- Allow for scheduling changing labels on a per-flow run basis - #109
- Fixes issue with Agent Config schema - #107
Released on September 24, 2020.
- Registering agents will now retrieve IDs from matching entries in order to prevent duplication - #95
- Add routes for setting run names - #77
- Fix
name
field indelete_agent
GraphQL resolver - #95 - Replace all uses of
database.models
withprefect.models
- #94
Released on September 15, 2020.
- Hasura metadata archives use the most recent alembic revision ID - #78
- Fix and consolidate behavior for active schedules - #86
Released on September 11, 2020.
- Add database structure and routes for persisting Agents - #58
- Add route for querying mapped children - #73
- Fix init.py for deprecated API module - #71
- Remove
run_count
andduration
columns - #72
- Increase performance of state update triggers - #72
Released on September 1, 2020.
- Automatically upgrade local database - #69
- Fix flow run creation not using default parameters from flow group - #64
- Ensure Lazarus ignores flow runs with active tasks - #67
- Deprecate flow settings that are actually flow group settings - #63
Released on August 26, 2020.
- Refactor Lazarus and Zombie Killer for modularity - #46
- Add API mode to queryable reference info - #59
Released on August 21, 2020.
- Upgrade Hasura to 1.3.0 - #18
- Set schedule to inactive if neither Flow nor Flow Group have a schedule - #37
- Dedupe flow group labels when set - #40
- Set Apollo's
post-start.sh
script to wait for the GraphQL service's healthcheck - #41 - Upgrade Hasura to 1.3.1 - #52
- Use Hasura default cache size - #52
- Adds RELEASE_TIMESTAMP for UI to display - #54
- Remove unused UI from services CLI - #45
- Cloud hooks require explicit states - #51
- Cloud hooks do not match state parents - #51
- Remove unused code and config settings - #53
- Remove unused
state_id
column - #35
- Release
- None
- None
- Fixes bug where updating a Flow Group schedule didn't clear old scheduled runs - #9
- None
- None