Runtime Metrics

Runtime metrics are available for runtime implementations that expose metric reporting.

The following tags are present on all runtime metrics:

The following tags are present on all tokio runtime metrics:

  • stat_id

  • runtime_id

    The runtime ID or label.

    There may be multiple instances of the same runtime type. This tag distinguishes between them.

Jemalloc Runtime Metrics

There metrics come from the jemalloc memory allocator.

  • runtime.jemalloc.approximate_active

    type: gauge

    The total number of bytes in active pages collected in an unsynchronized manner, without requiring an epoch update.

    Added in version 0.4.0.

Mimalloc Runtime Metrics

There metrics come from the mimalloc memory allocator.

  • runtime.mimalloc.current_pages

    type: gauge

    Current count of mimalloc pages.

  • runtime.mimalloc.peak_pages

    type: gauge

    Peak count of mimalloc pages.

  • runtime.mimalloc.current_commit

    type: gauge

    Current committed memory (backed by the page file).

  • runtime.mimalloc.peak_commit

    type: gauge

    Peak committed memory (backed by the page file).

Added in version 0.4.0.

Tokio Runtime Metrics

These metrics come from the Tokio runtime.

The supported metrics are:

  • runtime.tokio.alive_tasks

    type: gauge

    Current number of live tasks in the runtime.

  • runtime.tokio.global_queue_depth

    type: gauge

    Number of tasks currently scheduled in the runtime’s global queue.