A limited release of the platform including few improvements and bug-fixes while most of the current effort goes to new SEP (Its premiere is scheduled for the next release).

Improvements

• Progress on i3 Index optimization epic:
  • [FP-511] Review missing mongodb indexes and add them, e.g. one of them is for sure index on timestamp column
  • [FP-514] Optimize how documents' number calculation is performed due to some extremely long waiting time in some scenarios
• [FP-735] Use a DriverDataSource implementation instead of a HikariDataSource at i3 Connect / SpringJdbcDatabaseConfiguration
• [FP-736] Add default property for encryption key in signals / i3 Search
  • Every signal receiver on default has generated random encryption key which on SWARM/balanced installations with multiple search instances might ends with problems that payload generated in one search instance cannot be encoded in another search instance because different encryption keys. To lower risk of such problems, additional property has been added to application.properties and it's used on default in signal receivers instead of randomized value in constructor.
• [FP-787] Mitigate vulnerabilities in Solr and OpenSearch client libraries
  • CVE-2023-23612 - An issue how role claims are processed from the JWTs in OpenSearch, when the authentication backend is SAML or OpenID Connect
  • CVE-2023-34454 - Related to snappy-java (fast compressor/decompressor), due to unchecked multiplications, an integer overflow may cause an unrecoverable fatal error
• [FP-788] Mitigate vulnerabilities by upgrading thymeleaf to v3.1.2 and okhttp to v4.11.0
  • CVE-2023-34034 - Access Control Bypass in Spring Security
  • CVE-2023-34036 - Improper Neutralization of HTTP Headers for Scripting Syntax

Bug fixes

• [FP-342] When posting something incorrectly to Index, there’s no proper error information returned, so the response from Json document parser has been added and should help to investigate the reason why documents are failing
• [FP-755] List of searcher's features are sometimes only from CommonContext and those specific for search engine are missing
• [FP-775] Missing matches field in the search response, coming from the inner hits, while searching on a nested field
• [FP-777] Aggregation on non String fields is not possible (throws an exception Cannot get 'Sterms' variant: current variant is 'Lterms') when building a facet using ElasticFacetFeature

Removals

• None

Versions of key components

• Elasticsearch 7.17.9
• Apache Solr 8.11.2
• OpenSearch 1.3.12
• MongoDB 4.4.18
• Neo4j 4.4.11

Changelog