Merge pull request #4390 from esl/check-probe-metrics-types

Check probe metrics types

doc/operation-and-maintenance/MongooseIM-metrics.md

@@ -27,20 +27,20 @@ All metrics are divided into the following groups:
 === "Prometheus"
 
     <h3>`counter`</h3>
-    
+
     A monotonically increasing metric type. It is used for events like the number of stanzas processed by the system.
-    
+
     **Example:**
         ```
         # TYPE c2s_element_in_message_count counter
         # HELP c2s_element_in_message_count Event: c2s_element_in, Metric: message_count
         c2s_element_in_message_count{host_type="localhost"} 0
         ```
-    
+
     <h3>`gauge`</h3>
-    
+
     A metric that represents a current value in the system.
-    
+
     **Example:**
         ```
         # TYPE mnesia_info_running_db_nodes gauge
@@ -49,9 +49,9 @@ All metrics are divided into the following groups:
         ```
 
     <h3>`histogram`</h3>
-    
+
     A histogram collects values and groups them in buckets.
-    
+
     **Example:**
         ```
         # TYPE c2s_xmpp_element_size_in_byte_size histogram
@@ -61,35 +61,35 @@ All metrics are divided into the following groups:
         c2s_xmpp_element_size_in_byte_size_bucket{le="1073741824"} 0
         c2s_xmpp_element_size_in_byte_size_bucket{le="+Inf"} 0
         ```
-    
+
 === "Exometer"
-    
+
     <h3>`spiral`</h3>
-    
+
     This kind of metric provides 2 values: `total` event count (e.g. stanzas processed) and a value in 60s window (`one` value).
     Dividing `one` value by 60 provides an average per-second value over last minute.
-    
+
     **Example:** `[{total, 1000}, {one, 20}]`
-    
+
     <h3>`counter`</h3>
-    
+
     A simple monotonically increasing value. It consists of two properties:
-    
+
     **Example:** `[{value, 12}, {ms_since_reset, 91761}]`
-    
+
     <h3>`gauge`</h3>
-    
+
     It is similar to a `counter` type but can be set to any value.
-    
+
     * `value`
     * `ms_since_reset` - Time in milliseconds elapsed from the last metric update.
-    
+
     **Example:** `[{value, 12}, {ms_since_reset, 91761}]`
 
     <h3>`histogram`</h3>
-    
+
     A histogram collects values over a sliding window of 60s and exposes the following stats:
-    
+
     * `n` - A number of samples.
     * `mean` - An arithmetic mean.
     * `min`
@@ -228,26 +228,26 @@ Since Exometer doesn't support labels, the host types and tags are part of the m
     | Name | Type | Description (when it gets incremented) |
     | ---- | ---- | -------------------------------------- |
     | `wpool_rdbms_stats_workers`   | gauge | Number of workers in the pool |
-    | `wpool_rdbms_stats_recv_oct`  | counter  | Number of bytes received |
-    | `wpool_rdbms_stats_recv_cnt`  | counter  | Number of packets received |
-    | `wpool_rdbms_stats_recv_max`  | gauge   | Size of the largest packet, in bytes |
-    | `wpool_rdbms_stats_send_oct`  | counter  | Number of bytes sent |
-    | `wpool_rdbms_stats_send_max`  | gauge   | Size of the largest packet |
-    | `wpool_rdbms_stats_send_cnt`  | counter  | Number of packets sent |
-    | `wpool_rdbms_stats_send_pend` | counter  | Number of bytes waiting to be sent |
+    | `wpool_rdbms_stats_recv_oct`  | gauge | Number of bytes received |
+    | `wpool_rdbms_stats_recv_cnt`  | gauge | Number of packets received |
+    | `wpool_rdbms_stats_recv_max`  | gauge | Size of the largest packet, in bytes |
+    | `wpool_rdbms_stats_send_oct`  | gauge | Number of bytes sent |
+    | `wpool_rdbms_stats_send_max`  | gauge | Size of the largest packet |
+    | `wpool_rdbms_stats_send_cnt`  | gauge | Number of packets sent |
+    | `wpool_rdbms_stats_send_pend` | gauge | Number of bytes waiting to be sent |
 
 === "Exometer"
 
     | Name                                                         | Type    | Description (when it gets incremented) |
     |--------------------------------------------------------------|---------|----------------------------------------|
-    | `[HostType, wpool_rdbms_stats, PoolTag, workers]`   | counter | Number of workers in the pool          |
-    | `[HostType, wpool_rdbms_stats, PoolTag, recv_oct]`  | spiral  | Number of bytes received               |
-    | `[HostType, wpool_rdbms_stats, PoolTag, recv_cnt]`  | spiral  | Number of packets received             |
-    | `[HostType, wpool_rdbms_stats, PoolTag, recv_max]`  | gauge   | Size of the largest packet, in bytes   |
-    | `[HostType, wpool_rdbms_stats, PoolTag, send_oct]`  | spiral  | Number of bytes sent                   |
-    | `[HostType, wpool_rdbms_stats, PoolTag, send_max]`  | gauge   | Size of the largest packet             |
-    | `[HostType, wpool_rdbms_stats, PoolTag, send_cnt]`  | spiral  | Number of packets sent                 |
-    | `[HostType, wpool_rdbms_stats, PoolTag, send_pend]` | spiral  | Number of bytes waiting to be sent     |
+    | `[HostType, wpool_rdbms_stats, PoolTag, workers]`   | gauge | Number of workers in the pool          |
+    | `[HostType, wpool_rdbms_stats, PoolTag, recv_oct]`  | gauge | Number of bytes received               |
+    | `[HostType, wpool_rdbms_stats, PoolTag, recv_cnt]`  | gauge | Number of packets received             |
+    | `[HostType, wpool_rdbms_stats, PoolTag, recv_max]`  | gauge | Size of the largest packet, in bytes   |
+    | `[HostType, wpool_rdbms_stats, PoolTag, send_oct]`  | gauge | Number of bytes sent                   |
+    | `[HostType, wpool_rdbms_stats, PoolTag, send_max]`  | gauge | Size of the largest packet             |
+    | `[HostType, wpool_rdbms_stats, PoolTag, send_cnt]`  | gauge | Number of packets sent                 |
+    | `[HostType, wpool_rdbms_stats, PoolTag, send_pend]` | gauge | Number of bytes waiting to be sent     |
 
 When using a Rabbit worker pool, metrics defined in [mod_event_pusher_rabbit](../modules/mod_event_pusher_rabbit.md) are
 available.

src/instrument/mongoose_instrument.erl

@@ -46,6 +46,7 @@
 -type handler_key() :: atom(). % key in the `instrumentation' section of the config file
 -type handler_fun() :: fun((event_name(), labels(), config(), measurements()) -> any()).
 -type handlers() :: {[handler_fun()], config()}.
+-type handler_map() :: #{labels() => handlers()}.
 -type execution_time() :: integer().
 -type measure_fun(Result) :: fun((execution_time(), Result) -> measurements()).
 -type handler_module_opts() :: #{atom() => any()}.
@@ -274,16 +275,45 @@ set_up_and_register_event(EventName, Labels, Config, Events) ->
             {ExistingLabels, _, _} = maps:next(maps:iterator(HandlerMap)),
             case label_keys(ExistingLabels) of
                 LabelKeys ->
-                    Handlers = do_set_up(EventName, Labels, Config),
-                    Events#{EventName := HandlerMap#{Labels => Handlers}};
+                    handle_metrics_config(EventName, Labels, Config, HandlerMap, Events);
                 ExistingKeys ->
                     {error, #{what => inconsistent_labels,
                               event_name => EventName, labels => Labels,
                               existing_label_keys => ExistingKeys}}
             end;
         #{} ->
+            handle_metrics_config(EventName, Labels, Config, new, Events)
+    end.
+
+-spec handle_metrics_config(event_name(), labels(), config(), handler_map() | new, event_map()) ->
+          event_map() | {error, map()}.
+handle_metrics_config(EventName, Labels, Config, HandlerMap, Events) ->
+    case filter_improper_probe_metrics(Config) of
+        {ok, _} ->
             Handlers = do_set_up(EventName, Labels, Config),
-            Events#{EventName => #{Labels => Handlers}}
+            case HandlerMap of
+                new ->
+                    Events#{EventName => #{Labels => Handlers}};
+                _ ->
+                    Events#{EventName := HandlerMap#{Labels => Handlers}}
+            end;
+        {error, FilteredMetrics} ->
+            {error, #{what => non_gauge_metrics_in_probe,
+                      event_name => EventName, labels => Labels,
+                      improper_metrics => FilteredMetrics}}
+    end.
+
+-spec filter_improper_probe_metrics(config()) -> {ok, #{}} | {error, metrics()}.
+filter_improper_probe_metrics(Config) ->
+    case Config of
+        #{probe := _, metrics := Metrics} ->
+            FilteredMetrics = maps:filter(fun(_, V) -> V =/= gauge end, Metrics),
+            case map_size(FilteredMetrics) of
+                0 -> {ok, FilteredMetrics};
+                _ -> {error, FilteredMetrics}
+            end;
+        _ ->
+          {ok, #{}}
     end.
 
 -spec do_set_up(event_name(), labels(), config()) -> handlers().

src/wpool/mongoose_wpool_rdbms.erl

@@ -45,10 +45,10 @@ instrumentation(global, Tag) ->
     % Services use global pools. Since the same number of labels for a metric is expected, for an
     % event, global pool has to emit an event under a different name.
     [{wpool_global_rdbms_stats, #{pool_tag => Tag},
-      #{probe => #{module => ?MODULE}, metrics => rdbms_data_stats_measurement_types()}}];
+      #{probe => #{module => ?MODULE}, metrics => gauges([workers | inet_stats()])}}];
 instrumentation(HostType, Tag) ->
     [{wpool_rdbms_stats, #{host_type => HostType, pool_tag => Tag},
-      #{probe => #{module => ?MODULE}, metrics => rdbms_data_stats_measurement_types()}}].
+      #{probe => #{module => ?MODULE}, metrics => gauges([workers | inet_stats()])}}].
 
 -spec probe(mongoose_instrument:event_name(), mongoose_instrument:labels()) ->
     mongoose_instrument:measurements().
@@ -166,12 +166,6 @@ empty_inet_stats_measurements() ->
       send_cnt => 0,
       send_pend => 0}.
 
-rdbms_data_stats_measurement_types() ->
-    #{workers => counter,
-      recv_oct => spiral,
-      recv_cnt => spiral,
-      recv_max => gauge,
-      send_oct => spiral,
-      send_max => gauge,
-      send_cnt => spiral,
-      send_pend => spiral}.
+-spec gauges([mongoose_instrument:metric_name()]) -> mongoose_instrument:metrics().
+gauges(Keys) ->
+    maps:from_keys(Keys, gauge).

test/mongoose_instrument_SUITE.erl

@@ -37,6 +37,7 @@ api_test_cases() ->
      set_up_and_span_with_error,
      span_fails_when_not_set_up,
      set_up_probe,
+     set_up_probe_with_incorrect_metric_type,
      set_up_failing_probe,
      set_up_and_tear_down_probe,
      unexpected_events,
@@ -224,6 +225,19 @@ set_up_probe(Config) ->
                                              fun(L) -> length(L) > 1 end),
     ?assertEqual([ExpectedEvent, ExpectedEvent], History2).
 
+set_up_probe_with_incorrect_metric_type(Config) ->
+    Event = ?config(event, Config),
+    Cfg = #{metrics => #{test_metric => gauge},
+            probe => #{module => ?MODULE, interval => 1}},
+    Cfg1 = #{metrics => ImproperMetrics = #{test_metric => spiral},
+             probe => #{module => ?MODULE, interval => 1}},
+    Specs = [{Event, #{key => value1}, Cfg},
+             {Event, #{key => value2}, Cfg1}],
+    ?assertError(#{what := non_gauge_metrics_in_probe, improper_metrics := ImproperMetrics},
+                 mongoose_instrument:set_up(Event, ?LABELS, Cfg1)),
+    ?assertError(#{what := non_gauge_metrics_in_probe, improper_metrics := ImproperMetrics},
+                 mongoose_instrument:set_up(Specs)).
+
 set_up_failing_probe(Config) ->
     Event = ?config(event, Config),
     Labels = ?LABELS,