1.7.0

This release adds to the graphite plugin ability to send metrics
to the multiple servers and hide replication metrics of the deleted
replica.

Added

• graphite: ability to send metrics to the multiple servers.
  Backward compatibility with previous plugin version is preserved.
  From now on init method assigns an unique name to the created fiber
  using incoming graphite server opts (if passed). Added new stop()
  method to stop all fibers started by the plugin (#540).

Fixed

• Deleting a replica via box.space._cluster:delete() doesn't delete
  information about this replica from the metrics (it's gone only
  after cluster is restarted) (#538).

1.6.2

This release fixes the No such file or directory error that occurs
if the /proc/<pid>/stat file does not exist.

Fixed

• The error message level displayed when the /proc/<pid>/stat file is missing
  has been changed to verbose. (#536).

1.6.1

This release introduces fix of the box.schema.needs_upgrade metric.

Fixed

• Use box.schema.needs_upgrade instead of box.schema_needs_upgrade (#529).

1.6.0

This release adds new memory and schema metrics.

Added

• tnt_memory metric (#519).
• tnt_memory_virt metric (#521).
• tnt_schema_needs_upgrade metric (#524).

Fixed

• Fixed a potential process metrics error when fio.read returns an empty string or an error (#527).

1.5.0

Added

• tnt_cartridge_config_checksum metric.

Removed

• tnt_cartridge_config_applied metric that only reflected local node state.

1.4.0

Added

• tnt_cartridge_config_applied metric.
• New optional label_keys parameter for counter() and gauge() metrics.

1.3.1

Empty technical release. Identical to 1.3.0.

1.3.0

Added

• New metrics category in config: cpu_extended.

Fixed

• Use box.info.ro instead of box.cfg.read_only in replication metrics.

1.2.0

Added

• New Tarantool 3 metrics:
  • tnt_config_alerts
  • tnt_config_status

1.1.0

Added

• metrics.cfg{} "all" metasection for array include and exclude
  (metrics.cfg{include={'all'}} can be used instead of metrics.cfg{include='all'},
  metrics.cfg{exclude={'all'}} can be used instead of metrics.cfg{include='none'})

• tnt_election_leader_idle metric.

• Histogram now logs a warning if observe is called with cdata value.

Changed

• Inconsistent metrics descrtiptions for memtx metrics.

• New memory metrics:

  • tnt_memtx_tuples_data_total
  • tnt_memtx_tuples_data_read_view
  • tnt_memtx_tuples_data_garbage
  • tnt_memtx_index_total
  • tnt_memtx_index_read_view
  • tnt_vinyl_memory_tuple

Deprecated

• Using cdata values with histogram:observe.