dashboard: add syncro queue panels

Add panels to "Replication overview" section to dashboard templates.
Update consists of panels with following metrics:
- tnt_synchro_queue_owner
- tnt_synchro_queue_term
- tnt_synchro_queue_len
- tnt_synchro_queue_busy

Part of #178

Author: DifferentialOrange

CHANGELOG.md

@@ -11,6 +11,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 - Panels for transaction operations
 - Panels with net statistics per thread
 - Panel with failover trigger count
+- Panels with syncro replication queue statistics
 
 ### Changed
 - Replace LuaJit deprecated metrics with new ones

dashboard/panels/replication.libsonnet

@@ -137,4 +137,130 @@ local prometheus = grafana.prometheus;
       ).where('metric_name', '=', 'tnt_clock_delta').where('label_pairs_alias', '=~', alias)
       .selectField('value').addConverter('last')
   ),
+
+  local syncro_warning(description) = std.join(
+    '\n',
+    [description, |||
+      Panel works with metrics 0.15.0 or newer, Tarantool 2.8.1 or newer.
+    |||]
+  ),
+
+  synchro_queue_owner(
+    title='Synchronous queue owner',
+    description=syncro_warning(|||
+      Instance ID of the current synchronous replication master.
+    |||),
+    datasource_type=null,
+    datasource=null,
+    policy=null,
+    measurement=null,
+    job=null,
+    alias=null,
+  ):: common.default_graph(
+    title=title,
+    description=description,
+    datasource=datasource,
+    labelY1='owner id',
+    decimals=0,
+    panel_width=6,
+    legend_avg=false,
+    legend_max=false,
+  ).addTarget(common.default_metric_target(
+    datasource_type,
+    'tnt_synchro_queue_owner',
+    job,
+    policy,
+    measurement,
+    alias,
+  )),
+
+  synchro_queue_term(
+    title='Synchronous queue term',
+    description=syncro_warning(|||
+      Current queue term.
+    |||),
+    datasource_type=null,
+    datasource=null,
+    policy=null,
+    measurement=null,
+    job=null,
+    alias=null,
+  ):: common.default_graph(
+    title=title,
+    description=description,
+    datasource=datasource,
+    labelY1='term',
+    decimals=0,
+    panel_width=6,
+    legend_avg=false,
+    legend_max=false,
+  ).addTarget(common.default_metric_target(
+    datasource_type,
+    'tnt_synchro_queue_term',
+    job,
+    policy,
+    measurement,
+    alias,
+  )),
+
+  synchro_queue_length(
+    title='Synchronous queue transactions',
+    description=syncro_warning(|||
+      Count of transactions collecting confirmations now.
+    |||),
+    datasource_type=null,
+    datasource=null,
+    policy=null,
+    measurement=null,
+    job=null,
+    alias=null,
+  ):: common.default_graph(
+    title=title,
+    description=description,
+    datasource=datasource,
+    labelY1='current',
+    panel_width=6,
+  ).addTarget(common.default_metric_target(
+    datasource_type,
+    'tnt_synchro_queue_len',
+    job,
+    policy,
+    measurement,
+    alias,
+  )),
+
+  synchro_queue_busy(
+    title='Synchronous queue busy',
+    description=syncro_warning(|||
+      Whether the queue is processing any system entry (CONFIRM/ROLLBACK/PROMOTE/DEMOTE).
+
+      Panel works with Grafana 8.x.
+    |||),
+    datasource_type=null,
+    datasource=null,
+    policy=null,
+    measurement=null,
+    job=null,
+    alias=null,
+  ):: timeseries.new(
+    title=title,
+    description=description,
+    datasource=datasource,
+    panel_width=6,
+    max=1,
+    min=0,
+  ).addValueMapping(
+    1, 'yellow', 'busy'
+  ).addValueMapping(
+    0, 'green', 'not busy'
+  ).addRangeMapping(
+    0.001, 0.999, '-'
+  ).addTarget(common.default_metric_target(
+    datasource_type,
+    'tnt_synchro_queue_busy',
+    job,
+    policy,
+    measurement,
+    alias,
+  )),
 }

dashboard/section.libsonnet

@@ -165,6 +165,42 @@ local tdg_tuples = import 'dashboard/panels/tdg/tuples.libsonnet';
       job=job,
       alias=alias,
     ),
+
+    replication.synchro_queue_owner(
+      datasource_type=datasource_type,
+      datasource=datasource,
+      policy=policy,
+      measurement=measurement,
+      job=job,
+      alias=alias,
+    ),
+
+    replication.synchro_queue_term(
+      datasource_type=datasource_type,
+      datasource=datasource,
+      policy=policy,
+      measurement=measurement,
+      job=job,
+      alias=alias,
+    ),
+
+    replication.synchro_queue_length(
+      datasource_type=datasource_type,
+      datasource=datasource,
+      policy=policy,
+      measurement=measurement,
+      job=job,
+      alias=alias,
+    ),
+
+    replication.synchro_queue_busy(
+      datasource_type=datasource_type,
+      datasource=datasource,
+      policy=policy,
+      measurement=measurement,
+      job=job,
+      alias=alias,
+    ),
   ],
 
   http(datasource_type, datasource, policy=null, measurement=null, job=null, alias=null):: [

example_cluster/project/app/roles/storage.lua

@@ -1,5 +1,7 @@
 local function init(opts)
     if opts.is_master then
+        box.ctl.promote()
+
         for _, space_name in ipairs({'customers', 'clients'}) do
             local space = box.schema.space.create(space_name, {
                 format = {
@@ -10,6 +12,7 @@ local function init(opts)
                 },
                 if_not_exists = true,
                 engine = 'memtx',
+                is_sync = true,
             })
             -- primary index
             space:create_index('id_index', {

supported_metrics.md

@@ -121,10 +121,10 @@ Based on [tarantool/metrics 0.16.0](https://github.com/tarantool/metrics/release
 - [x] **tnt_cartridge_issues**: see *Cluster overview/Cartridge warning issues*, *Cluster overview/Cartridge critical issues* panels ([#55](https://github.com/tarantool/grafana-dashboard/pull/55))
 - **tnt_cartridge_cluster_issues**: unsupported (decided not to support: superseded by **tnt_cartridge_issues**)
 - [x] **tnt_cartridge_failover_trigger_total**: see *Cluster overview/Failovers triggered* panel ([#178](https://github.com/tarantool/grafana-dashboard/issues/178))
-- [ ] **tnt_synchro_queue_owner**: ([#178](https://github.com/tarantool/grafana-dashboard/issues/178))
-- [ ] **tnt_synchro_queue_term**: ([#178](https://github.com/tarantool/grafana-dashboard/issues/178))
-- [ ] **tnt_synchro_queue_len**: ([#178](https://github.com/tarantool/grafana-dashboard/issues/178))
-- [ ] **tnt_synchro_queue_busy**: ([#178](https://github.com/tarantool/grafana-dashboard/issues/178))
+- [x] **tnt_synchro_queue_owner**: see *Replication overview/Synchronous queue owner* panel ([#178](https://github.com/tarantool/grafana-dashboard/issues/178))
+- [x] **tnt_synchro_queue_term**: see *Replication overview/Synchronous queue term* panel ([#178](https://github.com/tarantool/grafana-dashboard/issues/178))
+- [x] **tnt_synchro_queue_len**: see *Replication overview/Synchronous queue transactions* panel ([#178](https://github.com/tarantool/grafana-dashboard/issues/178))
+- [x] **tnt_synchro_queue_busy**: see *Replication overview/Synchronous queue busy* panel ([#178](https://github.com/tarantool/grafana-dashboard/issues/178))
 - [ ] **tnt_election_state**: ([#178](https://github.com/tarantool/grafana-dashboard/issues/178))
 - [ ] **tnt_election_vote**: ([#178](https://github.com/tarantool/grafana-dashboard/issues/178))
 - [ ] **tnt_election_leader**: ([#178](https://github.com/tarantool/grafana-dashboard/issues/178))