* Significantly speed up recorder event listener
This code is called every time an event happens since it
subscribes to all events. Its our most frequently called
listener out of the box.
It used to have a seperate filter function but it was
later combined after core had some previous refactoring.
It was never optimized after that happened.
This change reduces the run time by ~70%
* decruft
* Ensure recorder run shutdown if the run loop raises
If anything goes wrong with the recorder we should
still try to shutdown cleanly
* tweak
* tests
* tests
* handle migraiton failure
* tweak comment
* naming
* order
* order
* order
* reword
* adjust test
* fixes
* threading
* failure case
* fix test
* have to wait for stop because the task blocks on thread join
* Reduce overhead of legacy database columns on new installs
* Reduce overhead of legacy database columns on new installs
* Reduce overhead of legacy database columns on new installs
* Reduce overhead of legacy database columns on new installs
* not working as expected
* override the type compiler
* override the type compiler
* override the type compiler
* override the type compiler
* Apply suggestions from code review
* pgsql char1
* make entity filter test setup with old schema
* fix some more tests that were mutating state
* fix some more tests that were mutating state
* fix some more tests that were mutating state
* fix more dbstate mutations
* add shim for older tests
* split migration tests
* add coverage for purging legacy data
* tweak
* more fixes
* drop some legacy
* fix another test
* fix a few more
* add casts for postgresql in case someone deletes the schema changes table
* dry
* dry
* dry
* Restart entity id post migration after a restart
If the entity migration finished and Home Assistant was
restarted during the post migration it would never be resumed
which means the old index and space would never be recovered
* add migration resume test
* Allow passing an optional name to async_track_time_interval
This is the same idea as passing a name to asyncio.create_task which
makes it easier to track down bugs
* more
* short
* still cannot find it
* add a few more
* test
* Fix cpu thrashing during purge after all legacy events were removed
We now remove the the index of of event ids on the states table when its
all NULLs to save space. The purge path needs to avoid checking for legacy
rows to purge if the index has been removed since it will result in a full
table scan each purge cycle that will always find no legacy rows to purge
* one more place
* drop the key constraint as well
* fixes
* more sqlite
* Avoid database executor job to fetch statistic metadata on cache hit
Since we will almost always have a cache hit fetching
statistic meta data we can avoid an executor job
* Avoid database executor job to fetch statistic metadata on cache hit
Since we will almost always have a cache hit fetching
statistic meta data we can avoid an executor job
* Avoid database executor job to fetch statistic metadata on cache hit
Since we will almost always have a cache hit fetching
statistic meta data we can avoid an executor job
* remove exception catch since the threading.excepthook will actually catch this in production
* fix a few missed ones
* threadsafe
* Update homeassistant/components/recorder/table_managers/statistics_meta.py
* coverage and optimistic caching
* refactor to make StatesMetaManager threadsafe
* refactor to make StatesMetaManager threadsafe
* refactor to make StatesMetaManager threadsafe
* refactor to make StatesMetaManager threadsafe
* reduce
* comments
* Split context id migration into states and events tasks
Since events can finish much earlier than states we
would keep looking at the table because states as not
done. Make them seperate tasks
* add retry dec
* fix migration happening twice
* another case
* Deduplicate event_types in the events table
* Deduplicate event_types in the events table
* more fixes
* adjust
* adjust
* fix product
* fix tests
* adjust
* migrate
* migrate
* migrate
* more test fixes
* more test fixes
* fix
* migration test
* adjust
* speed up
* fix index
* fix more tests
* handle db failure
* preload
* tweak
* adjust
* fix stale docs strings, remove dead code
* refactor
* fix slow tests
* coverage
* self join to resolve query performance
* fix typo
* no need for quiet
* no need to drop index already dropped
* remove index that will never be used
* drop index sooner as we no longer use it
* Revert "remove index that will never be used"
This reverts commit 461aad2c52.
* typo
* Load pending state attributes and event data ids at startup
Since we queue all events to be processed after startup
we can have a thundering herd of queries to prime the
LRUs of event data and state attributes ids. Since we
know we are about to process a chunk of events we can
fetch all the ids in two queries
* lru
* fix hang
* Fix recorder LRU being destroyed if event session is reopened
We would clear the LRU in _close_event_session but
it would never get replaced with an LRU again so
it would leak memory if the event session is reopened
* Fix recorder LRU being destroyed if event session is reopened
We would clear the LRU in _close_event_session but
it would never get replaced with an LRU again so
it would leak memory if the event session is reopened
* cleanup
* Adjust size of recorder LRU based on number of entities
If there are a large number of entities the cache would
get thrashed as there were more state attributes being
recorded than the size of the cache. This meant we had
to go back to the database to do lookups frequently when
an instance has more than 2048 entities that change
frequently
* add a test
* do not actually record 4096 states
* patch target
* patch target
* Chunk MariaDB data migration to avoid running out of buffer space
This will make the migration slower but since the innodb_buffer_pool_size
is using the defaul to 128M and not tuned to the db size there is a
risk of running out of buffer space for large databases
* Update homeassistant/components/recorder/migration.py
* hard code since bandit thinks its an injection
* Update homeassistant/components/recorder/migration.py
* guard against manually modified data/corrupt db
* adjust to 10k per chunk
* adjust to 50k per chunk
* memory still just fine at 250k
* but slower
* commit after each chunk to reduce lock pressure
* adjust
* set to 0 if null so we do not loop forever (this should only happen if the data is missing)
* set to 0 if null so we do not loop forever (this should only happen if the data is missing)
* tweak
* tweak
* limit cleanup
* lower limit to give some more buffer
* lower limit to give some more buffer
* where required for sqlite
* sqlite can wipe as many as needed with no limit
* limit on mysql only
* chunk postgres
* fix limit
* tweak
* fix reference
* fix
* tweak for ram
* postgres memory reduction
* defer cleanup
* fix
* same order
If the there are a lot of excluded events for the recorder, it
can have a performance impact as the list has to be searched
every time an event fires in HA
J. Nick Koston
epenet
Erik Montnemery
Joakim Plate