Tarantool 3 "replacements" for Cartridge

.luacheckrc

@@ -2,3 +2,4 @@ include_files = {"**/*.lua", "*.rockspec", "*.luacheckrc"}
 exclude_files = {"lua_modules/", ".luarocks/", ".rocks/", "tmp/", ".history/"}
 
 max_line_length = 120
+max_comment_line_length = 200

CHANGELOG.md

@@ -5,6 +5,10 @@ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
 ## [Unreleased]
+### Added
+- New Tarantool 3 metrics:
+  - tnt_config_alerts
+  - tnt_config_status
 
 ## [1.1.0] - 2024-05-17
 ### Added

doc/monitoring/api_reference.rst

@@ -566,6 +566,7 @@ Metrics functions
     *   ``cartridge_failover``
     *   ``clock``
     *   ``event_loop``
+    *   ``config``
 
     See :ref:`metrics reference <metrics-reference>` for details.
     All metric collectors from the collection have ``metainfo.default = true``.

doc/monitoring/metrics_reference.rst

@@ -993,3 +993,40 @@ Read view statistics
 
         *   -   ``tnt_memtx_index_read_view``
             -   Memory (in bytes) held for read views.
+
+
+Tarantool configuration
+-----------------------
+
+These metrics are available starting from Tarantool 3.0.
+
+..  container:: table
+
+    ..  list-table::
+        :widths: 25 75
+        :header-rows: 0
+
+        *   -   ``tnt_config_alerts``
+            -   Count of current instance :ref:`configuration apply alerts <config_api_reference_info>`.
+                ``{level="warn"}`` label covers warnings and
+                ``{level="error"}`` covers errors.
+
+        *   -   ``tnt_config_status``
+            -   The status of current instance :ref:`configuration apply <config_api_reference_info>`.
+                ``status`` label contains possible status name.
+                Current status has metric value ``1``, inactive statuses have metric value ``0``.
+
+                ..  code-block:: none
+
+                    # HELP tnt_config_status Tarantool 3 configuration status
+                    # TYPE tnt_config_status gauge
+                    tnt_config_status{status="reload_in_progress",alias="router-001-a"} 0
+                    tnt_config_status{status="uninitialized",alias="router-001-a"} 0
+                    tnt_config_status{status="check_warnings",alias="router-001-a"} 0
+                    tnt_config_status{status="ready",alias="router-001-a"} 1
+                    tnt_config_status{status="check_errors",alias="router-001-a"} 0
+                    tnt_config_status{status="startup_in_progress",alias="router-001-a"} 0
+
+                For example, this set of metrics means that current configuration
+                for ``router-001-a`` status is ``ready``.
+

metrics/tarantool.lua

@@ -23,6 +23,7 @@ local default_metrics = {
     cartridge_failover  = require('metrics.cartridge.failover'),
     clock               = require('metrics.tarantool.clock'),
     event_loop          = require('metrics.tarantool.event_loop'),
+    config              = require('metrics.tarantool.config'),
 }
 
 local all_metrics_map = {}

metrics/tarantool/config.lua

@@ -0,0 +1,76 @@
+local utils = require('metrics.utils')
+
+local collectors_list = {}
+
+local function get_config_alerts(config_info)
+    -- https://github.com/tarantool/tarantool/blob/319357d5973d15d08b8eda6a230eada08b710802/src/box/lua/config/utils/aboard.lua#L17-L18
+    local config_alerts = {
+        warn = 0,
+        error = 0,
+    }
+
+    for _, alert in pairs(config_info.alerts) do
+        config_alerts[alert.type] = config_alerts[alert.type] + 1
+    end
+
+    return config_alerts
+end
+
+local function get_config_status(config_info)
+    -- See state diagram here
+    -- https://github.com/tarantool/doc/issues/3544#issuecomment-1866033480
+    local config_status = {
+        uninitialized = 0,
+        startup_in_progress = 0,
+        reload_in_progress = 0,
+        check_warnings = 0,
+        check_errors = 0,
+        ready = 0,
+    }
+
+    config_status[config_info.status] = 1
+
+    return config_status
+end
+
+local function update()
+    if not utils.is_tarantool3() then
+        return
+    end
+
+    -- Can migrate to box.info().config later
+    -- https://github.com/tarantool/tarantool/commit/a1544d3bbc029c6fb2a148e580afe2b20e269b8d
+    local config = require('config')
+    local config_info = config:info()
+
+    local config_alerts = get_config_alerts(config_info)
+
+    for level, count in pairs(config_alerts) do
+        collectors_list.config_alerts = utils.set_gauge(
+            'config_alerts',
+            'Tarantool 3 configuration alerts',
+            count,
+            {level = level},
+            nil,
+            {default = true}
+        )
+    end
+
+    local config_status = get_config_status(config_info)
+
+    for status, value in pairs(config_status) do
+        collectors_list.config_status = utils.set_gauge(
+            'config_status',
+            'Tarantool 3 configuration status',
+            value,
+            {status = status},
+            nil,
+            {default = true}
+        )
+    end
+end
+
+return {
+    update = update,
+    list = collectors_list,
+}

metrics/utils.lua

@@ -36,4 +36,20 @@ function utils.delete_collectors(list)
     table.clear(list)
 end
 
+local function get_tarantool_version()
+    local version_parts = rawget(_G, '_TARANTOOL'):split('-', 3)
+
+    local major_minor_patch_parts = version_parts[1]:split('.', 2)
+    local major = tonumber(major_minor_patch_parts[1])
+    local minor = tonumber(major_minor_patch_parts[2])
+    local patch = tonumber(major_minor_patch_parts[3])
+
+    return major, minor, patch
+end
+
+function utils.is_tarantool3()
+    local major = get_tarantool_version()
+    return major == 3
+end
+
 return utils