1.5.5 (#198)

This release includes the following changes:

• @VictorErmakov updated the docker-compose sample with connection pool parameters to avoid fast connect cycling (#191).
• Update default values for connection pool parameters to use go-sql pooling by default to avoid fast connet cycling.
• Updated some third-party dependencies.
• Updated TxEventQ load generator utility.

Thank you to the following people for their suggestions and contributions:

@VictorErmakov

1.5.4

This release includes the following changes:

• Based of this recommendation from godror, which relates to the two following items, and in discussion with the ODPI-C team, we have introduced additional parameters to allow you to set connection pool parameters, and have set defaults which will avoid fast connect cycling. It is our expectation that a fix may be produced in the underlying ODPI-C library for the underlying issue. In the mean time, these changes will avoid the conditions under which the error can occur.
• Fix malloc error (#177, #181).
• Fix intermittent connection issues with ADB-S when exporter is run in a container (#169).
• Fix Multiple custom metrics files overwrite one another (#179).
• Replace go-kit/log with log/slog, due to upstream changes in prometheus/common.
• Add support for additional admin roles, expanding list of options for DB_ROLE to SYSDBA, SYSOPER, SYSBACKUP, SYSDG, SYSKM, SYSRAC and SYSASM (#180).
• Updated some third-party dependencies.

Thank you to the following people for their suggestions and contributions:

• @Jman1993
• @oey
• @jlembeck06
• @Jman1993
• @PeterP55P
• @rlagyu0
• @Sycri

Thank you to @tgulacsi for changes in godror (godror/godror#361, godror/godror#360), and to @cjbj and @sudarshan12s for support and guidance from ODPI-C (https://github.com/oracle/odpi).

In this release, we also continued some minor code refactoring.

1.5.3

Known issue: This release has a known issue that results in the error message malloc(): unsorted double linked list corrupted.
We recommend staying on 1.5.2 until a new release with a fix is available. We hope to have a fix by early March.

This release includes the following changes:

• Fix over-zealous supression of errors when ignorezeroresult = true (#168).
• When scrapeinterval is set, do first scrape immediately, not after the interval (#166).
• Updated some third-party dependencies.

Thank you to the following people for their suggestions and contributions:

• @redelang

In this release, we also started some minor code refactoring.

1.5.2

This release includes the following changes:

• Update the metric defintion for tablespace usage to report more accurate temp space usage.
• Revert InstantClient to 21c version due to ADB connectivity issue.
• Update documentation to explain how to obtain credentials from a wallet.
• Fix race condition on err variable in scrape() func (by @valrusu).
• Updated some third-party dependencies.

Thank you to the following people for their suggestions and contributions:

• @aureliocirella
• @mitoeth
• @valrusu

1.5.1

This release includes the following changes:

• Added support for using the TNS_ADMIN environment variable, which fixes an issue when connecting to Autonomous Database instances using TNS name.
• Updated InstantClient to 23ai version for amd64 and latest available 19.24 version for arm64.
• Fixed an issue with wrong LD_LIBRARY_PATH on some platforms. (#136)
• Added documentation and an example of using the scrapeinterval setting to change the interval at which a certain metric is colected.
• Added notes to documentation for extra security parameters needed when using a wallet with Podman.
• Updated some third-party dependencies.

Thank you for contributions from @AOzhogin, Michael and Toon.

1.5.0

This release includes the following changes:

• Support for running the exporter on ARM processors (darwin and linux).
• Updated some third-party dependencies.
• Updated the "test/demo environment" to use newer version of Oracle Database (23.5.0.24.07) and faster startup.

1.4.0

This release includes the following changes:

• Allow multiple custom metrics definition files.
• Allow query timeout per-metric.
• Allow scrape interval per-metric.
• Updated some third-party dependencies.

1.3.1

This release includes the following changes:

• Alert logs can be disabled by setting parameter log.disable to 1.
• Alert log exporter will stop if it gets three consecutive failures.
• Updated the list of required permissions.
• Updated the TxEventQ sample dashboard.
• Updated some third-party dependencies.

Thank you to the following people for their suggestions and contributions:

• @tux-jochen

1.3.0

This release includes the following changes:

• Alert logs can be exported for collection by a log reader like Promtail or FluentBit. Default
  output to /log/alert.log in JSON format.
• Provide ability to connect as SYSDBA or SYSOPER by setting DB_ROLE.
• New default metric is added to report the type of database connected to (CDB or PDB).
• New default metrics are added for cache hit ratios.
• Deafult metrics updated to suppress spurious warnings in log.
• Wait class metric updated to use a better query.
• The sample dashboard is updated to include new metrics.
• Fixed a bug which prevented periodic freeing of memory.
• Set CLIENT_INFO to a meaningful value.
• Update Go toolchain to 1.22.4.
• Updated some third-party dependencies.

Thank you to the following people for their suggestions and contributions:

• @pioro
• @savoir81

The updated sample dashboard looks like this:

1.2.1

This release includes the following changes:

• Accept max idle and open connections settings as parameters.
• Updated some third-party dependencies.