Timeseries Schemas

User scripts invoked via the user scripts hooks can use the Timeseries API to write their own timeseries data and then visualize it in Custom Pages. A full example of a plugin specialized in collecting Redis metrics is the Redis monitor plugin .

In order to write timeseries, it is first necessary to define the timeseries schemas.

Schemas Definition

Custom schemas are defined into the ./ts_schemas subdirectory of the plugin. The directory can contain one or more of the following files:

  • min.lua: Define schemas whose points have 1 minute resolution.
  • 5mins.lua: Define schemas whose points have 5 minutes resolution.
  • hour.lua: Define schemas whose points have 1 hour resolution.
  • day.lua: Define schemas whose points have 1 day resolution.

For example, schemas defined into min.lua are suitable to be used in the user scripts min hook (see the relevant page for more details). Each file should use the timeseries API in order to define its schemas. Here is an example of min.lua:

local ts_utils = require "ts_utils_core"
local schema

schema = ts_utils.newSchema("example:active_hosts", {
  step = 60,
  metrics_type = ts_utils.metrics.gauge,


The above script defines a schema named example:active_hosts whose points have a 60 seconds resolution. For internal reasons, it is always necessary to specify the step even if the file is named min.lua. The schema is identified by the ifid tag and contains one gauge metric named num_hosts.

Schemas Usage

A user script could then use the above schema as follows:

local script = {
  hooks = {},


script.hooks.min = function(params)

  if params.ts_enabled then
    local ifid = getSystemInterfaceId()
    local num = ... get the current metric value here ...

    ts_utils.append("example:active_hosts", {ifid = ifid, num_hosts = num})


It’s important to check that params.ts_enabled flag in order to call ts_utils.append only if the timeseries are enabled for the given entity currently processed.