Database system logs are a valuable source of data on the health of each infrastructure element. Due to the format being raw data, they are often not used. By compiling this information into a report, administrators have a more complete picture of the operation of the system. It is also a valuable source of information for data security teams.

PostgreSQL, being the heart of EuroDB, allows the logging of all relevant information describing the operation of a database cluster. EuroDB provides the analysis module for these logs. It compiles log data, presenting them in a graphical form. The information prepared in this way significantly reduces the time needed for analysis and detection of an anomaly or error.

Performance is the primary determinant in the construction of a log analyser. The parser was created to support large log files, including compressed (gzip) files.

Logs can be compiled at any interval thanks to the incremental mode. This allows for comparative analysis and storage of archived data.

EuroDB provides the following reports related to queries:

  • General statistics
  • Queries that are waiting most frequently
  • Queries with the longest waiting period
  • Queries that generate the largest number of temporary files
  • Queries that generate the largest temporary files
  • The slowest queries (longest running time)
  • Queries that took up the most time (quick in essence, whose execution took a lot of time to complete due to their complexity)
  • The most frequent queries
  • The most common errors
  • Users (at the database level) who generate the most queries
  • Most frequently cancelled queries

Generated graphs:

  • Lock statistics
  • Queries by type
  • Distribution of queries per database within a cluster
  • Sessions per database/user/client/application
  • Connections per database/user/client/application
  • Share of autovacuum and autoanalyse processes by tables
  • Query time histogram
  • Histogram of session times.

Statistics as a function of time:

  • Query statistics
  • Temporary file statistics
  • Checkpoint statistics
  • Autovacuum and autoanalyse statistics
  • Hourly histogram of the number and duration of queries
  • Hourly histogram of errors
  • Statistics on cancelled queries
  • Error statistics broken down by priority (PANIC/FATAL/ERROR/WARNING).