Document instrumentation

big_tests/tests/bosh_SUITE.erl

@@ -193,7 +193,7 @@ create_and_terminate_session(Config) ->
 
     % Assert that correct events have been executed
     [instrument_helper:assert(Event, Label, fun(#{byte_size := BS}) -> BS > 0 end)
-     || {Event, Label} <- instrumentation_events(), Event =/= c2s_message_processing_time],
+     || {Event, Label} <- instrumentation_events(), Event =/= c2s_message_processed],
 
     %% Verify C2S listener is not used
     instrument_helper:assert_not_emitted(negative_instrumentation_events()),
@@ -954,7 +954,7 @@ wait_for_zero_bosh_sessions() ->
 instrumentation_events() ->
     instrument_helper:declared_events(mod_bosh, [])
     ++ instrument_helper:declared_events(mongoose_c2s, [global])
-    ++ [{c2s_message_processing_time, #{host_type => host_type()}}].
+    ++ [{c2s_message_processed, #{host_type => host_type()}}].
 
 negative_instrumentation_events() ->
     [{Name, #{}} || Name <- negative_instrumentation_events_names()].

big_tests/tests/connect_SUITE.erl

@@ -407,7 +407,7 @@ metrics_test(Config) ->
     [instrument_helper:assert(Event, Label, fun(#{byte_size := BS}) -> BS > 0;
                                                (#{time := Time}) -> Time > 0 end)
      || {Event, Label} <- instrumentation_events(),
-        Event =/= c2s_message_processing_time].
+        Event =/= c2s_message_processed].
 
 tls_authenticate(Config) ->
     %% Given
@@ -818,4 +818,4 @@ proxy_info() ->
 instrumentation_events() ->
     instrument_helper:declared_events(mongoose_c2s_listener, [#{}])
     ++ instrument_helper:declared_events(mongoose_c2s, [global])
-    ++ [{c2s_message_processing_time, #{host_type => domain_helper:host_type()}}].
+    ++ [{c2s_message_processed, #{host_type => domain_helper:host_type()}}].

big_tests/tests/mim_c2s_SUITE.erl

@@ -241,7 +241,7 @@ escalus_start(Cfg, FlatCDs) ->
 instrumentation_events() ->
     instrument_helper:declared_events(mongoose_c2s_listener, [#{}])
     ++ instrument_helper:declared_events(mongoose_c2s, [global])
-    ++ [{c2s_message_processing_time, #{host_type => domain_helper:host_type()}}].
+    ++ [{c2s_message_processed, #{host_type => domain_helper:host_type()}}].
 
 tcp_instrumentation_events() ->
     [{c2s_tcp_data_out, #{}},
@@ -253,6 +253,6 @@ tls_instrumentation_events() ->
 
 common_instrumentation_events() ->
     HostType = domain_helper:host_type(),
-    [{c2s_message_processing_time, #{host_type => HostType}},
+    [{c2s_message_processed, #{host_type => HostType}},
      {c2s_xmpp_element_size_in, #{}},
      {c2s_xmpp_element_size_out, #{}}].

big_tests/tests/websockets_SUITE.erl

@@ -171,7 +171,7 @@ escape_attrs(Config) ->
 instrumentation_events() ->
     instrument_helper:declared_events(mod_websockets, [])
     ++ instrument_helper:declared_events(mongoose_c2s, [global])
-    ++ [{c2s_message_processing_time, #{host_type => domain_helper:host_type()}}].
+    ++ [{c2s_message_processed, #{host_type => domain_helper:host_type()}}].
 
 negative_instrumentation_events() ->
     [{Name, #{}} || Name <- negative_instrumentation_events_names()].

doc/configuration/instrumentation.md

@@ -0,0 +1,113 @@
+This section is used to configure MongooseIM instrumentation.
+It is a system of executing events when something of interest happens in the server.
+They are mainly used for the purpose of metrics.
+
+Instrumentation events are acted upon by handlers. Available instrumentation handlers are:
+
+* `prometheus` - exposes a metrics endpoint for [Prometheus](https://prometheus.io/).
+* `exometer` - starts [Exometer](https://github.com/esl/exometer_core), a metrics server capable of exporting metrics using reporters. Currently available is a [Graphite](https://graphiteapp.org/) reporter.
+* `log` - logs instrumentation events to disk.
+
+Enable them by adding a corresponding sections and possible configuration values.
+We recommend choosing either Prometheus or Exometer as a solution for exposing metrics.
+
+## General options
+
+### `instrumentation.probe_interval`
+* **Syntax:** positive integer
+* **Default:** `15` (seconds)
+* **Example:** `probe_interval = 60`
+
+Sets the interval for periodic measurements (probes).
+
+## Exometer options
+
+General options for the Exometer reporter:
+
+### `instrumentation.exometer.all_metrics_are_global`
+* **Syntax:** boolean
+* **Default:** `false`
+* **Example:** `all_metrics_are_global = true`
+
+## Exometer reporter options
+
+Multiple reporters can be configured.
+Because of that, each reporter is configured in a section inside a TOML array, for example: `[[instrumentation.exometer.report.graphite]]`.
+
+### `instrumentation.exometer.report.graphite.interval`
+* **Syntax:** positive integer
+* **Default:** `60000` (milliseconds)
+* **Example:** `interval = 30_000`
+
+Interval at which metrics will be sent to Graphite.
+
+### `instrumentation.exometer.report.graphite.host`
+* **Syntax:** string
+* **Default:** no default, required
+* **Example:** `host = "graphite.local"`
+
+The name or IP address of the Graphite server.
+This option is mandatory.
+
+### `instrumentation.exometer.report.graphite.port`
+* **Syntax:** integer, between 0 and 65535
+* **Default:** `2003`
+* **Example:** `port = 2033`
+
+The port on which the Graphite server listens for connections.
+
+### `instrumentation.exometer.report.graphite.connect_timeout`
+* **Syntax:** positive integer
+* **Default:** `5000` (milliseconds)
+* **Example:** `connect_timeout = 10_000`
+
+The amount of time Graphite reporter will wait before timing out.
+
+### `instrumentation.exometer.report.graphite.api_key`
+* **Syntax:** string
+* **Default:** `""`
+* **Example:** `api_key = "hosted_graphite_api_key"`
+
+API key to use when reporting to a hosted graphite server.
+
+### `instrumentation.exometer.report.graphite.prefix`
+* **Syntax:** string
+* **Default:** no default
+* **Example:** `prefix = "mim_stats"`
+
+A prefix to prepend all metric names with before they are sent to the graphite server.
+
+### `instrumentation.exometer.report.graphite.env_prefix`
+* **Syntax:** string
+* **Default:** no default
+* **Example:** `env_prefix = "GRAPHITE_METRICS_PREFIX"`
+
+Specifies an environmental variable name from which an additional prefix will be taken.
+In case both `prefix` and `env_prefix` are defined, it will be placed before the `prefix` and separated with a dot.
+
+## Example Prometheus configuration
+
+This configuration enables `prometheus`, and `log` handlers:
+```toml
+[instrumentation]
+  probe_interval = 10_000
+
+[instrumentation.prometheus]
+
+[instrumentation.log]
+```
+
+## Example Exometer configuration
+
+This configuration enables `exometer` handler with two different Graphite reporters.
+```toml
+[[instrumentation.exometer.report.graphite]]
+  host = "127.0.0.1"
+  interval = 15_000
+  prefix = "mongooseim"
+  connect_timeout = 5000
+
+[[instrumentation.exometer.report.graphite]]
+  host = "hosted_graphite.com"
+  prefix = "mim"
+```

doc/configuration/outgoing-connections.md

@@ -109,7 +109,7 @@ When MongooseIM fails to connect to the DB, it retries with an exponential backo
 * **Example:** `host = "localhost"`
 
 #### `outgoing_pools.rdbms.*.connection.port`
-* **Syntax:** string
+* **Syntax:** integer, between 0 and 65535
 * **Default:** `5432` for `pgsql`; `3306` for `mysql`
 * **Example:** `port = 5343`
 
@@ -204,7 +204,7 @@ There are two important limitations:
 * **Example:** `host = "redis.local"`
 
 ### `outgoing_pools.redis.*.connection.port`
-* **Syntax:** integer, between 0 and 65535, non-inclusive
+* **Syntax:** integer, between 0 and 65535
 * **Default:** `6379`
 * **Example:** `port = 9876`
 

doc/listeners/listen-http.md

@@ -13,6 +13,7 @@ Following configuration option is used to set up an HTTP handler:
 
     * `mod_bosh` - for [BOSH](https://xmpp.org/extensions/xep-0124.html) connections,
     * `mod_websockets` - for [WebSocket](https://tools.ietf.org/html/rfc6455) connections,
+    * `mongoose_prometheus_handler` - for [Prometheus]((https://prometheus.io/) metrics,
     * `mongoose_graphql_handler` - for GraphQL API,
     * `mongoose_admin_api`, `mongoose_client_api` - for REST API.
 
@@ -183,6 +184,12 @@ By default, all modules are enabled, so you don't need to change this option.
 The Swagger documentation of the client API is hosted at the `/api-docs` path.
 You can disable the hosted documentation by setting this option to `false`.
 
+## Handler types: Prometheus - `mongoose_prometheus_handler`
+
+Requires no additional options other than the [common handler options](#common-handler-options) in the listener section.
+In order to collect useful metrics, a `[prometheus]` section has to be added in [the instrumentation section](../configuration/instrumentation.md#).
+The default configuration available with MongooseIM is shown in [Example 7](#example-7-prometheus) below.
+
 ## Transport options
 
 The options listed below are used to modify the HTTP transport settings.
@@ -334,3 +341,18 @@ REST API for clients.
     host = "_"
     path = "/api"
 ```
+
+### Example 7. Prometheus
+
+Prometheus metrics endpoint.
+
+```toml
+[[listen.http]]
+  port = 9091
+
+  transport.num_acceptors = 10
+
+  [[listen.http.handlers.mongoose_prometheus_handler]]
+    host = "_"
+    path = "/metrics"
+```

doc/modules/mod_csi.md

@@ -25,7 +25,19 @@ Buffer size for messages queued when session was `inactive`.
 
 If you'd like to learn more about metrics in MongooseIM, please visit [MongooseIM metrics](../operation-and-maintenance/MongooseIM-metrics.md) page.
 
-| Name               | Type   | Description (when it gets incremented) |
-|--------------------|--------|----------------------------------------|
-| `mod_csi_active`   | spiral | A client becomes active.               |
-| `mod_csi_inactive` | spiral | A client becomes inactive.             |
+Prometheus metrics have a `host_type` label associated with these metrics.
+Since Exometer doesn't support labels, the host types, or word `global`, are part of the metric names, depending on the [`instrumentation.exometer.all_metrics_are_global`](../configuration/instrumentation.md#instrumentationexometerall_metrics_are_global) option.
+
+=== "Prometheus"
+
+    | Name | Type | Description (when it gets incremented) |
+    |------|------|----------------------------------------|
+    | `mod_csi_active_count` | counter | A client becomes active. |
+    | `mod_csi_inactive_count` | counter | A client becomes inactive. |
+
+=== "Exometer"
+
+    | Name | Type | Description (when it gets incremented) |
+    |------|------|----------------------------------------|
+    | `[HostType, mod_csi_active, count]` | spiral | A client becomes active. |
+    | `[HostType, mod_csi_inactive, count]` | spiral | A client becomes inactive. |

doc/modules/mod_event_pusher_http.md

@@ -110,10 +110,23 @@ Below is an example of what the body of an HTTP POST request can look like:
 
 If you'd like to learn more about metrics in MongooseIM, please visit [MongooseIM metrics](../operation-and-maintenance/MongooseIM-metrics.md) page.
 
-| Name | Type | Description (when it gets incremented) |
-| ---- | ---- | -------------------------------------- |
-| `[Host, mod_event_pusher_http, sent]` | spiral | An HTTP notification is sent successfully. |
-| `[Host, mod_event_pusher_http, failed]` | spiral | An HTTP notification failed. |
-| `[Host, mod_event_pusher_http, response_time]` | histogram | Does not include timings of failed requests. |
+Prometheus metrics have a `host_type` label associated with these metrics.
+Since Exometer doesn't support labels, the host types, or word `global`, are part of the metric names, depending on the [`instrumentation.exometer.all_metrics_are_global`](../configuration/instrumentation.md#instrumentationexometerall_metrics_are_global) option.
+
+=== "Prometheus"
+
+    | Name | Type | Description (when it gets incremented) |
+    | ---- | ---- | -------------------------------------- |
+    | `mod_event_pusher_http_sent_count` | counter | An HTTP notification is sent successfully. |
+    | `mod_event_pusher_http_sent_failure_count` | counter | An HTTP notification failed. |
+    | `mod_event_pusher_http_sent_response_time` | histogram | Time taken to send HTTP notification. Does not include timings of failed requests. |
+
+=== "Exometer"
+
+    | Name | Type | Description (when it gets incremented) |
+    | ---- | ---- | -------------------------------------- |
+    | `[Host, mod_event_pusher_http_sent, count]` | spiral | An HTTP notification is sent successfully. |
+    | `[Host, mod_event_pusher_http_sent, failure_count]` | spiral | An HTTP notification failed. |
+    | `[Host, mod_event_pusher_http_sent, response_time]` | histogram | Time taken to send HTTP notification. Does not include timings of failed requests. |
 
 [mod_event_pusher]: ./mod_event_pusher.md