metrics-lwt
  1. Overview
  2. Docs
On This Page
  1. Metric Monitoring
Legend:
Library
Module
Module type
Parameter
Class
Class type

Lwt monitoring.

The monitoring functions of this module return Lwt threads that proceed only when the metric operation is over, as defined by the current reporter.

v0.4.0 - homepage

Metric Monitoring

val add : ('a, 'b) Metrics.src -> ('a -> Metrics.tags) -> ('b -> Metrics.data Lwt.t) -> unit Lwt.t

add src t f adds a new data point to src.

val run : ('a, ('b, exn) result -> Metrics.Data.t) Metrics.src -> ('a -> Metrics.tags) -> (unit -> 'b Lwt.t) -> 'b Lwt.t

run src f runs f () and records in a new data point the time it took. run will also record the status of the computation, e.g. whether an exception has been raised.

val rrun : ('a, ('b, [ `Exn of exn | `Error of 'c ]) result -> Metrics.Data.t) Metrics.src -> ('a -> Metrics.tags) -> (unit -> ('b, 'c) result Lwt.t) -> ('b, 'c) result Lwt.t

Same as run but also record if the result is Ok or Error.

val periodically : (Metrics.field list, unit -> Metrics.data) Metrics.src -> unit

periodically src registers src for periodic collection.

val init_periodic : ?gc:[ `None | `Quick | `Full ] -> ?logs:bool -> (unit -> unit Lwt.t) -> unit

init_periodic ~gc ~logs sleeper starts a task which Lwt.join sleeper and all registered periodically sources. gc is by default `Full, collecting full GC stats - other options are `None and `Quick. If logs is provided and true (the default), the error and warning count from the Logs library are registered to be periodically reported.