Query logging

Since 4.1.0.

Query logging, disabled by default, prints one-line JSON objects into the standard output.

Its JSON format follows some ElasticSearch conventions and works nicely with the corresponding stack (FileBeat, ES, Kibana).

Example

{
  "@timestamp": "2023-10-13T14:02:05.973Z",
  "message": "query:all",
  "application": "ontop-odh",
  "payload": {
    "queryId": "36ade6c6-f091-4af2-9828-ee6316e191f4",
    "classesUsedInQuery": [
      "http://schema.org/Hotel"
    ],
    "propertiesUsedInQuery": [
      "http://schema.org/name"
    ],
    "tables": [
      "\"v_accommodationsopen\""
    ],
    "reformulationDuration": 50,
    "reformulationCacheHit": false,
    "user": "joe",
    "groups": "admins,users",
    "roles": "my_app_write,my_app_reader",
    "httpHeaders": {
      "referer": "http://localhost:8080/"
    },
    "extractedQueryTemplate": {
      "hash": "de167012e1ce3cf9dce66a986a3343d883a1f319f800132565b614498d9c73d8",
      "parameters": {
        "v0": "\"Residence Tuberis\""
      }
    },
    "sparqlQuery": "PREFIX schema: <http://schema.org/>\nSELECT * WHERE {\n  ?h a schema:Hotel ; schema:name ?name .\n  FILTER (langMatches(lang(?name), 'en'))\n  FILTER(str(?name) = \"Residence Tuberis\")\n} \nLIMIT 10",
    "reformulatedQuery": "ans1(h,name)\nCONSTRUCT [h, name] [name/\"Residence Tuberis\"^^@en, h/RDF(http://noi.example.org/data/accommodation/{}(VARCHARToTEXT(Id1m148)),IRI)]\n   NATIVE [Id1m148]\nSELECT v1.\"Id\" AS \"Id1m148\"\nFROM \"v_accommodationsopen\" v1\nWHERE ('Residence Tuberis' = v1.\"AccoDetail-en-Name\" AND 'HotelPension' = v1.\"AccoTypeId\")\nLIMIT 10\n",
    "executionBeforeUnblockingDuration": 8,
    "executionAndFetchingDuration": 30,
    "totalDuration": 80,
    "resultCount": 10
  }
}

Entries

Key	Type	Description
@timestamp	Timestamp
message	String	Type of the message. It always start with the prefix query:. Its default value is query:all. In case of an exception, the value starts with query:exception-.
application	String	Controlled by the property ontop.applicationName.
payload	JSON object	Gathers the entries below.
queryId	UUID	Unique to each query.
classesUsedInQuery	Array of IRIs	IRIs of the classes appearing in the SPARQL query.
propertiesUsedInQuery	Array of IRIs	IRIs of the properties appearing in the SPARQL query.
tables	Array of Strings	Names of the relations appearing in the SQL query.
reformulationDuration	Integer	Query reformulation duration (in ms).
reformulationCacheHit	Boolean	True if the reformulated query cache has been used
user	String	Since 5.2.0. User ID. Disabled by default
groups	String	Since 5.2.0. User groups separated by commas. Disabled by default
roles	String	Since 5.2.0. User roles separated by commas. Disabled by default
httpHeaders	JSON object	Values of a specific list of HTTP headers. By default, this list is empty. To track for instance the referer, set the property ontop.queryLogging.includeHttpHeader.referer to true.
extractedQueryTemplate. hash	String	Hash of the query template extracted, after removing extracted constants. Can be shared by similar queries.
extractedQueryTemplate. parameters	JSON object	Constants extracted and replaced by variables.
sparqlQuery	String	SPARQL query.
reformulatedQuery	String	Includes the SQL query and the post-processing.
executionBefore UnblockingDuration	Integer	Duration between the moments where the SQL query is sent and the result set is unblocked (in ms).
executionAndFetching Duration	Integer	Duration between the moments where the SQL query is sent and the last result is fetched (in ms).
totalDuration	Integer	Total duration (in ms).
resultCount	Integer	Number of results returned.

Default settings

# Query logging is disabled by default
ontop.queryLogging=false
# Application name (needed for logging)
ontop.applicationName=Ontop
# Includes the SPARQL query string into the query log
ontop.queryLogging.includeSparqlQuery=true
# Includes the reformulated query into the query log
ontop.queryLogging.includeReformulatedQuery=false
# Includes classes and properties into the query log
ontop.queryLogging.includeClassesAndProperties=true
# Includes DB tables/views into the query log
ontop.queryLogging.includeTables=true
# Includes the user ID, his/her groups and roles
ontop.queryLogging.includeUserInfo=false
# Provides separated message at different phases (after reformulation, result set unblocked, last result fetched)
ontop.queryLogging.decomposition=false
# Sets that merged messages are only inserted when decomposition is disabled
ontop.queryLogging.decompositionAndMergingMutuallyExclusive=true