INC-143121 · Issue 610732
Timeout for loading predictors made configurable
Resolved in Pega Version 8.3.6
When using an extremely large number of predictors, the Report definition pzADMPredictorsFilter was suffering timeouts due to the time for loading predictors from the database exceeding the time threshold allowed. This has been resolved by marking the rule as editable to allow custom setting of the threshold according to need.
INC-148899 · Issue 615702
Adaptive models update correctly
Resolved in Pega Version 8.3.6
Some models had the recorded responses column updated, but the models (number of Positive, Negative and Processed Responses) were not updated. Investigation showed that deleting the modelRuleConfiguration through the stateManager/client did not delete modelFactories related to the configuration. If a new configuration came in with a different algorithm, the update issue occurred. This has been resolved by reseting the configuration according to its factory in that specific case.
INC-155822 · Issue 618267
Locking added to avoid null pointer error for auto-populate property
Resolved in Pega Version 8.3.6
After configuring the auto populate property "OrgProduct" which referred to a data page, the system experiencing heavy load led to the property not getting properly initialized. This resulted in a WrongModeException and NullPointerException. To resolve this, the system has been updated to lock the requestor when Queue Processors execute their activity. This will prevent race conditions and concurrent modifications if other threads are accessing the same requestor.
INC-157629 · Issue 626633
Duplicate key exception resolved for adaptive model
Resolved in Pega Version 8.3.6
During the model snapshot update, a DuplicateKeyException was generated while trying to insert a record in to the predictor table. This did not affect the model's learning, but did appear ion the model monitoring report. This was traced to a local scenario of having the same outcome values defined on the model with different cases (Accept and accept). All predictors used in an Adaptive model are inserted into the model monitoring tables as a part of the monitoring job: because the monitoring tables are not case sensitive, this lead to a unique constraint exception since there were multiple IH predictors with the same name. To resolve this, validation has been added which will skip adding duplicates from new responses.
INC-160331 · Issue 628710
ML model continue tag error changed to debug logging
Resolved in Pega Version 8.3.6
Log entries were seen indicating "The continue tag is found without a start tag for window Line1 1234567890 in text". This appeared to be coming from the MLCommand java class. Investigation showed this occurred when the ML model believed a portion of the text (token) was a continuation of the entity because it has not found the starting part of the entity. This situation is not currently treated as a valid output, so an update has been made to change the logging level from error to debug for this case.
SR-69015 · Issue 619995
Unescaping characters implemented for expressions
Resolved in Pega Version 8.3.6, Resolved in Pega Version 8.4.4, Resolved in Pega Version 8.5.3, Resolved in Pega Version 8.6
An issue where expression builder statements were evaluated differently at runtime than at testing has been resolved. Pega Platform expressions with String literals(that is, sequences of characters enclosed in quotation marks) now unescape characters in strategy shapes such as Set Property or Filter.
SR-D90367 · Issue 556687
Cleanup enhanced for long pyEditElement names
Resolved in Pega Version 8.5
A pyEditElement error relating to decision data was seen multiple times in a stack trace. Research showed that while the utility worked as expected for decision data rules with names of less than 30 characters, the pyEditElement section was truncated the name for the decision data. This meant that decision data with the name SampleIssueandSampleGroupTwosalkdjkightntbmkblffvfvfv would be saved as SampleIssueandSampleGroupT for the pyEditElement section. Because of this, the utility failed the match and did not clean up the pyEditElement section. To resolve this, the cleanup utility has been updated to handle pyEditElement sections of decision data with longer names. Additional logging has also been added to improve debugging.
SR-D71621 · Issue 533296
Real time processing picks up correct datetime for Capture Response records
Resolved in Pega Version 8.5
A Realtime Data flow for the Capture Response flow was configured with a strategy shape set to load previous decisions within the past 7 days. Once this Realtime DF was started, attempting to Capture Response for decisions made after that startup timepoint did not work. This was traced to the InteractionID being written with global properties for the datetimes, and has been resolved by making those datetime properties local so the start and end time are not cached and the time range is calculated based on "now”.
SR-D85558 · Issue 548286
Handling added for prolonged Heartbeat Update Queries
Resolved in Pega Version 8.5
After restart, the pyFTSIncrementalIndexer queue size had hundreds of thousands of entries even though it was empty prior to the restart. Investigation traced this to a job scheduler that checked all the database connections everyday at 1 EST by using a list that contained some connections which did not exist. Checking those invalid connections caused other update queries to queue and wait, resulting in the update heartbeat query taking longer than its default beat. This caused a Split Brain issue wherein other nodes considered the long-executing node to be dead and triggered a rebalance while the node itself continued to execute partitions thinking that it was healthy. This caused duplicate processing of records. To resolve this, a fail safe has been added: while updating heartbeat in Service Registry, nodes will enter safe mode when the update query is taking longer than the default beat.
SR-D66397 · Issue 530333
ADM out-of-sync corrected for multi-datacenter Cassandra cluster
Resolved in Pega Version 8.5
After setting up the multi-datacenter configuration for a Cassandra cluster that consisted of six nodes in datacenter 1 and three nodes in datacenter 2, failover testing revealed a mismatch in the number of ADM models stored in each datacenter. The mismatch was observed mostly in the number of records present in the "adm_scoringmodel" and "adm_response_commit_log_date_tiered" tables. When Cassandra nodes are down, the other nodes in the cluster will store hints (records to be written) for the down nodes. When these nodes come back online the hints are replayed to those nodes and the data is written. Hints are written for 3 hours, so if a node come back up within 3 hours data is recovered and repairs are not required. The gc_grace_seconds for the above tables that were getting out of sync across the two datacenters was set to zero seconds. The "gc_grace_seconds" attribute is not just used as the time for removal of tombstones, it's also used to set the TTL for records written to the system.hints table. That meant that when the hints were written for the ADM tables for the nodes that were down, they were immediately expired since it was set to 0 and not played back when the terminated nodes restarted and joined the cluster. This has been resolved with this fix for all customers new to this release. Existing customers already on v7.3 or higher will need to complete the local change detailed below: Connect to the Cassandra cluster using cqlsh in the Pega Cassandra distribution and then run ALTER TABLE adm_commitlog.adm_response_commit_log_date_tiered WITH gc_grace_seconds = 86400; to change the relevant setting from zero to the equivalent of one day - the same length of time that the data in the table lives for. This will mean that any hints written can still be used to replay data to another node while the data itself is alive. It does also mean, however, that, given a constant load, a day's worth of expired ADM event data in the table will always be present on the disk, as the tombstones can now not be cleaned up for a day.