gitlab-exporter 17.0.1 → 17.1.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 15b0fdb50984859832e6a21dcb76c31dfbfb68f28e5601f2c1389782ea4b5f2f
-  data.tar.gz: a02d8bda949c2d5f94981d50123ff845e73d958306d0bcf9e3ca0cfd4c8107e5
+  metadata.gz: df145039780458bab34b436ca149dce37c4c4f4a08e1707b606f70dc4aa021a4
+  data.tar.gz: 272f4e78ffb5901954f086e58a158087328d0478deb18114fb748504e067796b
 SHA512:
-  metadata.gz: ab40fc6975a31c6175cb3139dd1e576a06f7f1a6b598efc387f76228e68413a47b7902d7aca1682128ee77dd76ee4f2e178fec7bed8dc027676ec2e614289490
-  data.tar.gz: e749cdfd3a8db77f8a594ad7007db5dc919003defa0afb7c3691a461b78d0bcdf459e8bce6d5106809f0cdc567323a968fe1d8de88c990a2c5074339b6c38343
+  metadata.gz: a223ec3d899fcc85eeb391e100cce3942aeb050264cc03c6877ad6387f09234461408f45f1a8a82128c657b240f4ce2768c35a1d58857d4c3fb212f42090f28c
+  data.tar.gz: 992f12485329239412a793b67c556c7226535765d3ecfea83271e055c679c75d7d1d0af8b0be71e83408a0aacc77ceb0e3d9079281725d26b04f86955e01baf7

data/Gemfile.lock

@@ -1,11 +1,11 @@
 PATH
   remote: .
   specs:
-    gitlab-exporter (17.0.1)
+    gitlab-exporter (17.1.0)
       base64 (= 0.3.0)
       connection_pool (= 2.5.5)
       deep_merge (~> 1.2.2)
-      faraday (= 2.14.3)
+      faraday (= 2.14.4)
       pg (= 1.6.3)
       quantile (= 0.2.1)
       redis (= 4.8.1)
@@ -21,7 +21,7 @@ GEM
     connection_pool (2.5.5)
     deep_merge (1.2.2)
     diff-lcs (1.5.0)
-    faraday (2.14.3)
+    faraday (2.14.4)
       faraday-net_http (>= 2.0, < 3.5)
       json
       logger
@@ -57,8 +57,8 @@ GEM
       diff-lcs (>= 1.2.0, < 2.0)
       rspec-support (~> 3.12.0)
     rspec-support (3.12.0)
-    rubocop (1.89.0)
-      json (~> 2.3)
+    rubocop (1.91.0)
+      json (>= 2.3)
       language_server-protocol (~> 3.17.0.2)
       lint_roller (~> 1.1.0)
       parallel (>= 1.10)
@@ -91,4 +91,4 @@ DEPENDENCIES
   rubocop (~> 1.0)
 
 BUNDLED WITH
-  4.0.10
+  4.0.18

data/README.md

@@ -23,14 +23,20 @@ metrics.
      `gitlab_pg_sequences_min_value`, `gitlab_pg_sequences_max_value`, `gitlab_pg_sequences_current_value`
     * [Row count queries](lib/gitlab_exporter/database/row_count.rb) --
       `gitlab_database_rows`
-    * [CI builds](lib/gitlab_exporter/database/ci_builds.rb) --
+    * [CI builds](lib/gitlab_exporter/database/ci_builds.rb) (**deprecated**) --
       `ci_pending_builds`, `ci_created_builds`, `ci_stale_builds`,
-      `ci_running_builds`
+      `ci_running_builds`, `ci_unarchived_traces`
+      (see [Deprecated: CI builds metrics](#deprecated-ci-builds-metrics))
     * [Bloat](lib/gitlab_exporter/database/bloat.rb) --
       `gitlab_database_bloat_$type_$key` with type `btree` (index bloat) or `table`
       (table bloat) and keys `bloat_ratio bloat_size extra_size real_size` (see below)
     * [Remote mirrors](lib/gitlab_exporter/database/remote_mirrors.rb) --
       `project_remote_mirror_last_successful_update_time_seconds`, `project_remote_mirror_last_update_time_seconds`
+    * [Geo node status](lib/gitlab_exporter/database/geo_node_status.rb) -- Geo metrics from
+      `geo_node_statuses`, emitted as `geo_<status_field>` gauges with trailing `_count` and
+      `_total` suffixes removed, matching Rails' exported metric fields and `name`/`url` labels.
+      Usage-data counters are excluded. Enable the `geo_node_status` probe in the `database`
+      endpoint configuration.
     * [Zoekt](lib/gitlab_exporter/database/zoekt.rb) --
       `search_zoekt_task_processing_queue_size`
 1. Git
@@ -133,6 +139,57 @@ Also see the [original documentation](https://github.com/ioguix/pgsql-bloat-esti
 Note that all metrics returned are estimates without an upper bound for
 the error.
 
+### Deprecated: CI builds metrics
+
+The [CI builds probe](lib/gitlab_exporter/database/ci_builds.rb) is deprecated and
+will be removed in a future major version. It collects its metrics by querying
+`p_ci_builds` directly, which puts avoidable load on tables that job scheduling
+already loads heavily, and it has had almost no maintenance since 2021. GitLab
+Rails and GitLab Runner instrument most of the same information by counting
+events as they happen; GitLab.com stopped using this probe for CI state years
+ago.
+
+The probe prints a deprecation notice on every run, and each of its metrics
+carries `(deprecated, ...)` in its `# HELP` line, so the warning reaches
+Prometheus and Grafana and not only the exporter's log.
+
+Replacements, where one exists:
+
+| Deprecated metric | Replacement |
+| --- | --- |
+| `ci_pending_builds` | `gitlab_ci_current_queue_size` (Rails, gauge, by `runner_type`) |
+| `ci_running_builds` | `gitlab_runner_jobs` (Runner, gauge, by `runner`, `state`, `stage`) |
+| `ci_created_builds` | None. A `created` build is not queued yet, so nothing counts it. |
+| `ci_stale_builds` | None. The nearest signal is `gitlab_ci_queue_operations_total{operation="build_status_stale"}` (Rails, counter), which counts stale rows met while queueing rather than how many are stale now. |
+| `ci_unarchived_traces` | None. |
+
+The replacements are not a like-for-like relabel: the Rails queue metrics are
+keyed by `runner_type` and the Runner metrics by runner, so neither carries the
+`namespace` label this probe reports. Queue health is better served by the Rails
+metrics that have no exporter equivalent at all -- `job_queue_duration_seconds`
+(how long jobs wait), `job_register_attempts_total` and
+`job_register_attempts_failed_total`, and `gitlab_ci_queue_operations_total`.
+Those four are always instrumented.
+
+**`gitlab_ci_current_queue_size`, and every other `gitlab_ci_queue_*` metric, is
+behind the `gitlab_ci_builds_queuing_metrics` ops feature flag, which is disabled
+by default.** Enable it before pointing a dashboard at them, or the replacement
+panels stay empty:
+
+```ruby
+Feature.enable(:gitlab_ci_builds_queuing_metrics)
+```
+
+Metric definitions live in
+[`lib/gitlab/ci/queue/metrics.rb`](https://gitlab.com/gitlab-org/gitlab/-/blob/master/lib/gitlab/ci/queue/metrics.rb)
+(Rails) and
+[`commands/builds_helper.go`](https://gitlab.com/gitlab-org/gitlab-runner/-/blob/main/commands/builds_helper.go)
+(Runner). See
+[GitLab Prometheus metrics](https://docs.gitlab.com/administration/monitoring/prometheus/gitlab_metrics/)
+for enabling the Rails exporter, and
+[Monitoring runners](https://docs.gitlab.com/runner/monitoring/) for the
+Runner's.
+
 ## Contributing
 
 gitlab-exporter is an open source project and we are very happy to accept community contributions. Please refer to [CONTRIBUTING.md](/CONTRIBUTING.md) for details.

data/config/gitlab-exporter.yml.example

@@ -77,6 +77,9 @@ probes:
         <<: *db_common_opts
         project_ids:
         - 1
+      geo_node_status:
+        class_name: Database::GeoNodeStatusProber
+        <<: *db_common
     zoekt:
       class_name: Database::ZoektProber
       <<: *db_common

data/gitlab-exporter.gemspec

@@ -28,7 +28,7 @@ Gem::Specification.new do |s|
   s.add_runtime_dependency "base64", "0.3.0"
   s.add_runtime_dependency "connection_pool", "2.5.5"
   s.add_runtime_dependency "deep_merge", "~> 1.2.2"
-  s.add_runtime_dependency "faraday", "2.14.3"
+  s.add_runtime_dependency "faraday", "2.14.4"
   s.add_runtime_dependency "pg", "1.6.3"
   s.add_runtime_dependency "quantile", "0.2.1"
   s.add_runtime_dependency "redis", "4.8.1"

data/lib/gitlab_exporter/database/ci_builds.rb

@@ -294,12 +294,53 @@ module GitLab
 
       # The prober which is called when gathering metrics
       class CiBuildsProber
+        # The stdout notice in #probe_db reaches whoever reads the exporter's log,
+        # which is not who reads the metrics. HELP is the one channel that travels
+        # with the sample to the scraper, so the deprecation rides there too: it
+        # survives into Prometheus and shows up in Grafana's metric browser.
+        DEPRECATED = "(deprecated, to be removed in a future major version)".freeze
+        REPLACEMENT = "See the CI builds entry in README.md for replacements.".freeze
+
+        PrometheusMetrics.describe(
+          "ci_created_builds",
+          "#{DEPRECATED} Number of CI builds in the created state, by namespace. #{REPLACEMENT}",
+          "gauge"
+        )
+
+        PrometheusMetrics.describe(
+          "ci_pending_builds",
+          "#{DEPRECATED} Number of CI builds in the pending state, by namespace. #{REPLACEMENT}",
+          "gauge"
+        )
+
+        PrometheusMetrics.describe(
+          "ci_stale_builds",
+          "#{DEPRECATED} Number of running CI builds not updated for over an hour. #{REPLACEMENT}",
+          "gauge"
+        )
+
+        PrometheusMetrics.describe(
+          "ci_running_builds",
+          "#{DEPRECATED} Number of running CI builds, by runner and namespace. #{REPLACEMENT}",
+          "gauge"
+        )
+
+        PrometheusMetrics.describe(
+          "ci_unarchived_traces",
+          "#{DEPRECATED} Number of finished CI builds whose trace is not archived. #{REPLACEMENT}",
+          "gauge"
+        )
+
         def initialize(metrics: PrometheusMetrics.new, **opts)
           @metrics = metrics
           @collector = CiBuildsCollector.new(**opts)
         end
 
         def probe_db
+          puts "[DEPRECATED] probe_db and CiBuildsProber are now considered obsolete"\
+            " and will be removed in a future major version,"\
+            " please use the CI/CD queuing metrics emitted by GitLab Rails instead"
+
           @results = @collector.run
 
           ci_builds_metrics(@results[:created_builds], "ci_created_builds") if @results[:created_builds]

data/lib/gitlab_exporter/database/geo_node_status.rb

@@ -0,0 +1,116 @@
+module GitLab
+  module Exporter
+    module Database
+      # A helper class to collect Geo node status metrics.
+      class GeoNodeStatusCollector < Base
+        QUERY = <<~SQL.freeze
+          SELECT
+            geo_nodes.name AS geo_node_name,
+            geo_node_statuses.db_replication_lag_seconds,
+            geo_node_statuses.replication_slots_count,
+            geo_node_statuses.replication_slots_used_count,
+            geo_node_statuses.replication_slots_max_retained_wal_bytes,
+            geo_node_statuses.last_event_id,
+            geo_node_statuses.cursor_last_event_id,
+            geo_node_statuses.status,
+            EXTRACT(EPOCH FROM geo_node_statuses.last_event_date)::bigint AS last_event_timestamp,
+            EXTRACT(EPOCH FROM geo_node_statuses.cursor_last_event_date)::bigint AS cursor_last_event_timestamp,
+            EXTRACT(EPOCH FROM geo_node_statuses.last_successful_status_check_at)::bigint
+              AS last_successful_status_check_timestamp
+          FROM geo_node_statuses
+          INNER JOIN geo_nodes ON geo_nodes.id = geo_node_statuses.geo_node_id
+          WHERE geo_nodes.enabled = TRUE
+        SQL
+
+        def run
+          with_connection_pool do |conn|
+            conn.exec(QUERY)
+          end
+        rescue PG::UndefinedTable, PG::UndefinedColumn
+          nil
+        end
+      end
+
+      # The prober which is called when gathering metrics.
+      class GeoNodeStatusProber
+        METRIC_PREFIX = "geo_".freeze
+        COUNT_SUFFIX = /_(count|total)\z/
+        METRIC_ALIASES = { "project_repositories_count" => "repositories_count" }.freeze
+        COLUMN_METRICS = %w[
+          db_replication_lag_seconds
+          replication_slots_count
+          replication_slots_used_count
+          replication_slots_max_retained_wal_bytes
+          last_event_id
+          cursor_last_event_id
+        ].freeze
+        TIMESTAMP_METRICS = %w[
+          last_event_timestamp
+          cursor_last_event_timestamp
+          last_successful_status_check_timestamp
+        ].freeze
+        REPLICATOR_METRIC_SUFFIXES = %w[
+          _count
+          _checksum_total_count
+          _checksummed_count
+          _checksum_failed_count
+          _synced_count
+          _failed_count
+          _registry_count
+          _verification_total_count
+          _verified_count
+          _verification_failed_count
+          _oldest_unsynced_time
+        ].freeze
+        NON_PROMETHEUS_STATUS_FIELDS = %w[projects_count].freeze
+
+        def initialize(metrics: PrometheusMetrics.new, **opts)
+          @metrics = metrics
+          @collector = opts[:collector] || GeoNodeStatusCollector.new(**opts)
+        end
+
+        def probe_db
+          @collector.run.to_a.each do |row|
+            labels = {
+              name: row["geo_node_name"],
+              url: row["geo_node_name"]
+            }
+
+            row.fetch("status", {}).to_h.each do |name, value|
+              add_metric(name, value, labels) if value.is_a?(Integer) && prometheus_metric?(name)
+            end
+
+            (COLUMN_METRICS + TIMESTAMP_METRICS).each do |name|
+              add_metric(name, row[name], labels) if row[name].is_a?(Integer)
+            end
+          end
+
+          self
+        rescue PG::ConnectionBad
+          self
+        end
+
+        def write_to(target)
+          target.write(@metrics.to_s)
+        end
+
+        private
+
+        def prometheus_metric?(name)
+          name = name.to_s
+          return false if NON_PROMETHEUS_STATUS_FIELDS.include?(name)
+          return false if COLUMN_METRICS.include?(name)
+
+          REPLICATOR_METRIC_SUFFIXES.any? { |suffix| name.end_with?(suffix) }
+        end
+
+        def add_metric(name, value, labels)
+          name = METRIC_ALIASES.fetch(name.to_s, name.to_s)
+          metric_name = "#{METRIC_PREFIX}#{name.sub(COUNT_SUFFIX, '')}"
+          PrometheusMetrics.describe(metric_name, "Geo node status metric", "gauge")
+          @metrics.add(metric_name, value, **labels)
+        end
+      end
+    end
+  end
+end

data/lib/gitlab_exporter/database.rb

@@ -8,6 +8,7 @@ module GitLab
       autoload :RowCountProber,       "gitlab_exporter/database/row_count"
       autoload :BloatProber,          "gitlab_exporter/database/bloat"
       autoload :RemoteMirrorsProber,  "gitlab_exporter/database/remote_mirrors"
+      autoload :GeoNodeStatusProber,  "gitlab_exporter/database/geo_node_status"
       autoload :PgSequencesProber,    "gitlab_exporter/database/pg_sequences"
       autoload :ZoektProber,          "gitlab_exporter/database/zoekt"
     end

data/lib/gitlab_exporter/version.rb

@@ -1,5 +1,5 @@
 module GitLab
   module Exporter
-    VERSION = "17.0.1".freeze
+    VERSION = "17.1.0".freeze
   end
 end

data/spec/database/ci_builds_spec.rb

@@ -361,3 +361,55 @@ describe GitLab::Exporter::Database do
     end
   end
 end
+
+# Asserting on the notice keeps a later refactor of #probe_db from dropping it
+# silently, which would leave operators unwarned before the probe goes (issue #119).
+describe GitLab::Exporter::Database::CiBuildsProber do
+  let(:prober) do
+    described_class.new(connection_string: "host=localhost",
+                        metrics: GitLab::Exporter::PrometheusMetrics.new(include_timestamp: false))
+  end
+
+  let(:metric_names) do
+    %w[ci_created_builds ci_pending_builds ci_stale_builds ci_running_builds ci_unarchived_traces]
+  end
+
+  it "warns that the probe is deprecated and names the replacement" do
+    allow_any_instance_of(GitLab::Exporter::Database::CiBuildsCollector).to receive(:run)
+      .and_return(pending_builds: [], stale_builds: 0, per_runner: [], unarchived_traces: 0)
+
+    expect { prober.probe_db }.to output(/\[DEPRECATED\].*CiBuildsProber.*GitLab Rails/m).to_stdout
+  end
+
+  # The stdout notice never reaches the scraper. HELP does, so every metric the
+  # probe can emit has to carry the deprecation, not just the ones with a
+  # replacement.
+  it "describes every metric it can emit as deprecated" do
+    metric_names.each do |name|
+      expect(GitLab::Exporter::PrometheusMetrics.description(name))
+        .to start_with("(deprecated"), "#{name} is emitted without a deprecated HELP line"
+      expect(GitLab::Exporter::PrometheusMetrics.metric_type(name)).to eq("gauge")
+    end
+  end
+
+  it "renders the deprecation into the scrape output" do
+    allow_any_instance_of(GitLab::Exporter::Database::CiBuildsCollector).to receive(:run)
+      .and_return(pending_builds: [{ namespace: "1", shared_runners: "yes", value: 12 }],
+                  stale_builds: 3, per_runner: [], unarchived_traces: 0)
+
+    output = StringIO.new
+    capture_stdout do prober.probe_db.write_to(output) end
+
+    expect(output.string).to include("# HELP ci_pending_builds (deprecated")
+    expect(output.string).to include("# TYPE ci_pending_builds gauge")
+    expect(output.string).to include(%(ci_pending_builds{namespace="1",shared_runners="yes"} 12.0))
+  end
+
+  def capture_stdout
+    original = $stdout
+    $stdout = StringIO.new
+    yield
+  ensure
+    $stdout = original
+  end
+end

data/spec/database/geo_node_status_spec.rb

@@ -0,0 +1,210 @@
+require "spec_helper"
+require "gitlab_exporter/database/geo_node_status"
+
+describe GitLab::Exporter::Database::GeoNodeStatusCollector do
+  let(:connection_pool) { double("connection pool") }
+  let(:connection) { double("connection") }
+  let(:results) { [{ "geo_node_name" => "secondary", "status" => { "projects_count" => 12 } }] }
+
+  subject(:collector) { described_class.new(connection_string: "host=localhost") }
+
+  before do
+    allow(collector).to receive(:connection_pool).and_return(connection_pool)
+    allow(connection_pool).to receive(:with).and_yield(connection)
+  end
+
+  describe "QUERY" do
+    it "selects all required status columns, timestamps, and node name" do
+      expect(described_class::QUERY).to include("geo_nodes.name AS geo_node_name")
+      expect(described_class::QUERY).to include("geo_node_statuses.db_replication_lag_seconds")
+      expect(described_class::QUERY).to include("geo_node_statuses.replication_slots_count")
+      expect(described_class::QUERY).to include("geo_node_statuses.replication_slots_used_count")
+      expect(described_class::QUERY).to include("geo_node_statuses.replication_slots_max_retained_wal_bytes")
+      expect(described_class::QUERY).to include("geo_node_statuses.last_event_id")
+      expect(described_class::QUERY).to include("geo_node_statuses.cursor_last_event_id")
+      expect(described_class::QUERY).to include("geo_node_statuses.status")
+      expect(described_class::QUERY).to include(
+        "EXTRACT(EPOCH FROM geo_node_statuses.last_event_date)::bigint AS last_event_timestamp"
+      )
+      expect(described_class::QUERY).to include(
+        "EXTRACT(EPOCH FROM geo_node_statuses.cursor_last_event_date)::bigint AS cursor_last_event_timestamp"
+      )
+      expect(described_class::QUERY).to include(
+        "EXTRACT(EPOCH FROM geo_node_statuses.last_successful_status_check_at)::bigint"
+      )
+    end
+  end
+
+  describe "#run" do
+    it "queries enabled Geo nodes and returns their status data" do
+      expect(connection).to receive(:exec).with(described_class::QUERY).and_return(results)
+
+      expect(collector.run).to eq(results)
+    end
+
+    it "returns nil when Geo tables are unavailable" do
+      allow(connection).to receive(:exec).and_raise(PG::UndefinedTable)
+
+      expect(collector.run).to be_nil
+    end
+  end
+end
+
+describe GitLab::Exporter::Database::GeoNodeStatusProber do
+  let(:writer) { StringIO.new }
+  let(:metrics) { GitLab::Exporter::PrometheusMetrics.new(include_timestamp: false) }
+  let(:collector) { double(GitLab::Exporter::Database::GeoNodeStatusCollector, run: results) }
+  let(:results) do
+    [
+      {
+        "geo_node_id" => 7,
+        "geo_node_name" => "secondary",
+        "db_replication_lag_seconds" => 4,
+        "replication_slots_count" => 2,
+        "replication_slots_used_count" => 1,
+        "replication_slots_max_retained_wal_bytes" => 1024,
+        "last_event_id" => 100,
+        "cursor_last_event_id" => 98,
+        "last_event_timestamp" => 1_750_000_100,
+        "cursor_last_event_timestamp" => 1_750_000_090,
+        "last_successful_status_check_timestamp" => 1_750_000_000,
+        "status" => {
+          "projects_count" => 12,
+          "project_repositories_count" => 21,
+          "projects_verification_total_count" => 8,
+          "projects_synced_count" => 19,
+          "projects_failed_count" => 2,
+          "git_fetch_event_count_weekly" => 100,
+          "container_repositories_replication_enabled" => true,
+          "status_message" => "Healthy"
+        }
+      }
+    ]
+  end
+
+  subject(:prober) { described_class.new(metrics: metrics, collector: collector) }
+
+  describe "constants and data sources" do
+    it "defines the expected COLUMN_METRICS corresponding to table columns" do
+      expect(described_class::COLUMN_METRICS).to contain_exactly(
+        "db_replication_lag_seconds",
+        "replication_slots_count",
+        "replication_slots_used_count",
+        "replication_slots_max_retained_wal_bytes",
+        "last_event_id",
+        "cursor_last_event_id"
+      )
+    end
+
+    it "defines the expected TIMESTAMP_METRICS corresponding to epoch extractions" do
+      expect(described_class::TIMESTAMP_METRICS).to contain_exactly(
+        "last_event_timestamp",
+        "cursor_last_event_timestamp",
+        "last_successful_status_check_timestamp"
+      )
+    end
+
+    it "does not define CORE_STATUS_METRICS" do
+      expect(described_class.const_defined?(:CORE_STATUS_METRICS)).to be false
+    end
+
+    it "does not emit column metrics if placed only inside status jsonb" do
+      allow(collector).to receive(:run).and_return([
+                                                     {
+                                                       "geo_node_name" => "secondary",
+                                                       "status" => {
+                                                         "db_replication_lag_seconds" => 4,
+                                                         "replication_slots_count" => 2,
+                                                         "replication_slots_used_count" => 1,
+                                                         "replication_slots_max_retained_wal_bytes" => 1024,
+                                                         "last_event_id" => 100,
+                                                         "cursor_last_event_id" => 98
+                                                       }
+                                                     }
+                                                   ])
+
+      prober.probe_db
+      prober.write_to(writer)
+
+      expect(writer.string).not_to include("geo_db_replication_lag_seconds")
+      expect(writer.string).not_to include("geo_replication_slots")
+      expect(writer.string).not_to include("geo_replication_slots_used")
+      expect(writer.string).not_to include("geo_replication_slots_max_retained_wal_bytes")
+      expect(writer.string).not_to include("geo_last_event_id")
+      expect(writer.string).not_to include("geo_cursor_last_event_id")
+    end
+
+    it "does not emit replicator metrics if placed only in row columns outside status jsonb" do
+      allow(collector).to receive(:run).and_return([
+                                                     {
+                                                       "geo_node_name" => "secondary",
+                                                       "project_repositories_count" => 21,
+                                                       "projects_synced_count" => 19,
+                                                       "status" => {}
+                                                     }
+                                                   ])
+
+      prober.probe_db
+      prober.write_to(writer)
+
+      expect(writer.string).not_to include("geo_repositories")
+      expect(writer.string).not_to include("geo_projects_synced")
+    end
+  end
+
+  describe "#probe_db" do
+    it "exports integer Geo metrics with Rails-compatible names and site labels" do
+      prober.probe_db
+      prober.write_to(writer)
+      labels = 'name="secondary",url="secondary"'
+
+      expect(writer.string).to include("geo_db_replication_lag_seconds{#{labels}} 4")
+      expect(writer.string).to include("# TYPE geo_db_replication_lag_seconds gauge")
+      expect(writer.string).to include("geo_replication_slots{#{labels}} 2")
+      expect(writer.string).to include("# TYPE geo_replication_slots gauge")
+      expect(writer.string).to include("geo_replication_slots_used{#{labels}} 1")
+      expect(writer.string).to include("# TYPE geo_replication_slots_used gauge")
+      expect(writer.string).to include("geo_replication_slots_max_retained_wal_bytes{#{labels}} 1024")
+      expect(writer.string).to include("# TYPE geo_replication_slots_max_retained_wal_bytes gauge")
+      expect(writer.string).to include("geo_last_event_id{#{labels}} 100")
+      expect(writer.string).to include("geo_cursor_last_event_id{#{labels}} 98")
+      expect(writer.string).to include("geo_last_event_timestamp{#{labels}} 1750000100")
+      expect(writer.string).to include("geo_cursor_last_event_timestamp{#{labels}} 1750000090")
+      expect(writer.string).to include("geo_last_successful_status_check_timestamp{#{labels}} 1750000000")
+      expect(writer.string).to include("geo_repositories{#{labels}} 21")
+      expect(writer.string).to include("geo_projects_verification_total{#{labels}} 8")
+      expect(writer.string).to include("geo_projects_synced{#{labels}} 19")
+      expect(writer.string).to include("geo_projects_failed{#{labels}} 2")
+
+      expect(writer.string).not_to include("event_log_max_id")
+      expect(writer.string).not_to include("geo_projects{")
+      expect(writer.string).not_to include("geo_git_fetch_event")
+      expect(writer.string).not_to include("container_repositories_replication_enabled")
+      expect(writer.string).not_to include("status_message")
+    end
+
+    it "handles nil column values gracefully" do
+      allow(collector).to receive(:run).and_return([
+                                                     {
+                                                       "geo_node_name" => "secondary",
+                                                       "db_replication_lag_seconds" => nil,
+                                                       "replication_slots_count" => nil,
+                                                       "status" => { "project_repositories_count" => 5 }
+                                                     }
+                                                   ])
+
+      prober.probe_db
+      prober.write_to(writer)
+
+      expect(writer.string).not_to include("geo_db_replication_lag_seconds")
+      expect(writer.string).not_to include("geo_replication_slots")
+      expect(writer.string).to include('geo_repositories{name="secondary",url="secondary"} 5')
+    end
+
+    it "handles a missing status payload" do
+      allow(collector).to receive(:run).and_return([{ "geo_node_name" => "secondary" }])
+
+      expect { prober.probe_db }.not_to raise_error
+    end
+  end
+end

data/spec/prometheus_metrics_spec.rb

@@ -29,15 +29,29 @@ describe GitLab::Exporter::PrometheusMetrics do
 
     allow(Time).to receive(:now).and_return(time)
 
-    described_class.describe("mymetric", "description")
-    described_class.describe("missingmetric", "otherdescription")
-    subject.add("mymetric", 1.3, mylabel: "x", myotherlabel: "y")
-
-    expect(subject.to_s).to eq(<<~METRICS)
+    # Descriptions are class-level state, and probes register theirs when their
+    # file loads, which happens once. Describing on a subclass keeps this example
+    # out of the shared registry: clearing it here would drop those registrations
+    # for every later example, and nothing can put them back.
+    metrics_class = Class.new(described_class)
+    metrics = metrics_class.new
+
+    metrics_class.describe("mymetric", "description")
+    metrics_class.describe("missingmetric", "otherdescription")
+    metrics.add("mymetric", 1.3, mylabel: "x", myotherlabel: "y")
+
+    expect(metrics.to_s).to eq(<<~METRICS)
       # HELP mymetric description
       mymetric{mylabel="x",myotherlabel="y"} 1.3 #{(time.to_f * 1000).to_i}
     METRICS
+  end
+
+  it "clears descriptions" do
+    metrics_class = Class.new(described_class)
+    metrics_class.describe("mymetric", "description", "gauge")
 
-    described_class.clear_descriptions
+    expect { metrics_class.clear_descriptions }
+      .to change { metrics_class.description("mymetric") }.from("description").to(nil)
+    expect(metrics_class.metric_type("mymetric")).to be_nil
   end
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: gitlab-exporter
 version: !ruby/object:Gem::Version
-  version: 17.0.1
+  version: 17.1.0
 platform: ruby
 authors:
 - Pablo Carranza
@@ -58,14 +58,14 @@ dependencies:
     requirements:
     - - '='
       - !ruby/object:Gem::Version
-        version: 2.14.3
+        version: 2.14.4
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - '='
       - !ruby/object:Gem::Version
-        version: 2.14.3
+        version: 2.14.4
 - !ruby/object:Gem::Dependency
   name: pg
   requirement: !ruby/object:Gem::Requirement
@@ -209,6 +209,7 @@ files:
 - lib/gitlab_exporter/database/bloat_btree.sql
 - lib/gitlab_exporter/database/bloat_table.sql
 - lib/gitlab_exporter/database/ci_builds.rb
+- lib/gitlab_exporter/database/geo_node_status.rb
 - lib/gitlab_exporter/database/pg_sequences.rb
 - lib/gitlab_exporter/database/remote_mirrors.rb
 - lib/gitlab_exporter/database/row_count.rb
@@ -230,6 +231,7 @@ files:
 - spec/cli_spec.rb
 - spec/database/bloat_spec.rb
 - spec/database/ci_builds_spec.rb
+- spec/database/geo_node_status_spec.rb
 - spec/database/pg_sequences_spec.rb
 - spec/database/row_count_spec.rb
 - spec/database/tuple_stats_spec.rb
@@ -274,6 +276,7 @@ test_files:
 - spec/cli_spec.rb
 - spec/database/bloat_spec.rb
 - spec/database/ci_builds_spec.rb
+- spec/database/geo_node_status_spec.rb
 - spec/database/pg_sequences_spec.rb
 - spec/database/row_count_spec.rb
 - spec/database/tuple_stats_spec.rb