AbstractTbQueueConsumerTemplate.poll() returned emptyList() immediately
when partitions was empty, bypassing both doPoll() and the secondary
sleep guard (which is also skipped for backends that report
isLongPollingSupported()==true, e.g. Kafka). The result was a
permanent CPU-burning loop on consumers whose partition assignment
ended up empty after a rebalance cascade — observed on 26
ie-downlink-consumer threads (~244% total CPU) until container restart.
Route the empty-partition path through sleepAndReturnEmpty() so the
caller honors durationInMillis regardless of long-polling support.
The alreadyProcessed() method used separate get() and put() calls on the
local cache, allowing concurrent threads in the notification executor pool
to both read null and bypass deduplication, creating duplicate notifications.
Replace with a single compute() call that atomically checks and updates
the cache entry, preventing the race between concurrent trigger processing.
Also fix: discard external cache timestamps that are more than 1 hour in
the future (clock skew protection), and avoid reading back from the SOFT
ref local cache when writing to external cache (GC could null it out).
* Experiments with CSV
* CSV Loader v1
* EDQ tests
* Volatile variables instead of final
* Improvements
* updated loader with new entities
* Fix double memory usage issue
* Basic data structures and load
* Minor improvements
* Snappy + Large String reuse
* added EntityFields classes for each entity
* Basic implementation
* Minor improvements to KeyFilters
* implemented RepositoryUtils.checkKeyFilters
* Generic query implementation
* New structure
* Refactoring and few processors implementation
* extended DeviceData with shared/client attributes and device profile
* Minor refactoring of attribute scopes
* DeviceTypeFilter support
* Strong types of fields for each entity data class
* DeviceType and AssetType filters
* EntityView and Edge queries
* Relations Query
* Relation Query Implementation
* Update EDQS module version
* Sync with EDQS via Kafka
* EDQS: major refactoring
* EDQS API requests via Kafka
* EDQS: full sync with the database
* Refactoring for EDQS sync
* EDQS: major refactoring and new features
* EDQS refactoring, count query support, fix tests
* EDQS: refactoring for query processors
* Fix EDQS pom version
* Cleanup edqs.yml
* EDQS: tenant partitioning strategy; refactoring
* EDQS: latest events queue
* EDQS: support for monolith setup; RocksDB; other improvements
* EDQS: merge sync and events topics, introduce state topic
* EDQS: dynamic repartitioning
* implemented entity data query filters for edqs
* EdqsEntityQueryControllerTest - use in-memory queue
* edqs-filter fixes, added test
* EDQS: blob entity support
* EdqsEntityQueryControllerTest - use in-memory queue
* Use DummyEdqsService when disabled
* Fixes for EDQS
* Refactoring for EDQS tests
* Fix edqs requests partitioning
* EDQS: Fix for attributes handling
* Fix attributes saving in EntityServiceTest
* EDQS: refactoring, fixes
* Minor refactoring for query processor
* added ownerName/ownerType support
* fixed relation query processor
* fixed EntityServiceTest
* refactoring
* added support for parentId for relation query result
* Get rid of EntityNameFetcher
* Add fixme for relation query processor
* db restore with select all edqs fields
* fixed entity deletion
* fixed FieldUtils with new EntityFields
* dao method renamed
* EDQS: instance groups with same partitions; automatic sync; multiple fixes
* Refactoring for EDQS sync
* EDQS: refactoring
* Fix startup with Kafka
* fixed EntityQueryControllerTest
* fixed EdqsEntityServiceTest
* Separate queue admin for EDQS request template
* Implement new EDQS partitioning strategy
* EDQS: multiple fixes and refactoring
* Add mock EdqsRocksDb beans to tests
* added edqs stats for inmemory/grafana
* fixed filter tests
* Update todos
* Refactoring for QueueConfig
* Improvements and refactoring for EDQS consumers
* implemented TODOs
* test fixes
* Consume state topic up to end offsets
* edqs stats refactoring
* EDQS: cleanup on partitions removal; refactoring
* EDQS: minor refactoring
* EDQS: remove CSV loader
---------
Co-authored-by: Andrii Shvaika <ashvayka@thingsboard.io>
Co-authored-by: dashevchenko <dshevchenko@thingsboard.io>
Local Subscription Service refactoring
Refactoring of the proto structures
Bug fixing after refactoring
Minor improvements and bug fixes
Improvements to avoid additional DB calls
Bug fixes after refactoring
Cluster mode improvements
Volodymyr Babak
Sergey Matvienko
Viacheslav Klimov
Andrii Landiak
Andrii Shvaika
Igor Kulikov
Dmytro Skarzhynets
ThingsBoard Bot
Andrii Landiak
Oleksandra Matviienko
Ivan Raznatovskyi
YevhenBondarenko
driesva
Seraphym-Tuhai
ShvaykaD