wget -O hdxctl https://hdx-infrastructure.s3.amazonaws.com/hdxctl-v2.22.6 && chmod +x hdxctl
- Support for AWS ap-southeast-2 and ap-northeast-1
- Replace Amazon Load Balancers with Nginx and NLB's
- Move Load balancers to the Core side of the stack.
- Publish native protocol port for Query access on load balancers, now able to use DBeaver type tools
Configuration and Control
- Improve availability of docker repos.
- Increase IAM Policy granularity
- Update to the
hdxctl scaleto have
--minimalto bring up 1 instance of Grafana, Superset and Prometheus
hdxctl instancescommand for improved instance definition.
hdxproject name is now a reserved name and can not be used.
- Add support for schema-less AVRO support.
- Improve Merge Logging
- Improve and reorganize stream peer shutdown sequence for faster and more reliable shutdown.
- Stream Rejects rows on failure instead of whole batches when limits are hit.
- Add ability within transform to specify more than one potential input source field.
- Collect internal Grafana metrics in Prometheus
- Grafana World Map plugin installation
- Grafana Tree Map plugin installation
- Grafana Diagram plugin installation
- Expose all
catalogcolumn metadata on tables.
- Add version endpoint accessible via NGINX
- Add NGINX metrics to Prometheus
- Add AWS t2 family server type to
- Change API definition to OpenAPIv3 format with various improvements.
- Add Servers field to the OpenAPIv3 definition for better documentation support.
- Keycloak performance improvements.
- Portal: Add drawer to Users page in the UI.
- Portal: Add error messages to Job detail drawer.
- query.timeout_secs - Introduce a query timeout circuit breaker Query API HTTP Options
- storage.read_max_partitions - Maximum number of partitions to be read for a query circuit breaker Query API HTTP Options
- storage.max_timerange_secs - Maximum time range a query can span circuit breaker Query API HTTP Options
- Improve connection handling and reliability when a peer is removed or added to a query pool.
Configuration and Control
- Bugfix for Manifest syncing failures when null resolution types are experienced.
- HDXCTL: Fix silently failing bootstrap template build
- Fix missing permissions for cached docker images that could affect deployment and update.
- Fix issue causing the need to run
routetwice when updating a cluster.
- Fix connection errors on webs instances during rolling restarts
- Bug fix for new client stacks not getting DNS entries
- Bug fix for old ClientID's, where old ClientID's could not be re-used in a new deployment
- Fix stream double counting of some CSV-in-JSON rows.
- Resolves an issue where Batch Peer can deadlock and take no more work after certain failure modes
- Constrain Journald data to 1G
- Limit size of container logs to 10M for instances running Podman.
- Allow for null "resolution" in non datetime types in transform definition.
- Various Portal functionality fixes
- Various networking and reliability Bug Fixes.