Closes #694 - Added option to expose counters as differences to influ…

…xDB exporter (#695)

Co-authored-by: Marius Oehler <[email protected]>

.gitignore

@@ -31,3 +31,4 @@ hooks
 /.nb-gradle/
 
 /working_directory/
+inspectit-ocelot-demo/configuration-server/users.db

components/inspectit-ocelot-eum-server/build.gradle

@@ -61,7 +61,7 @@ dependencies {
 			'com.maxmind.geoip2:geoip2:2.12.0',
 			'commons-net:commons-net:3.3',
 			'org.springframework.boot:spring-boot-starter-actuator',
-            "rocks.inspectit:opencensus-influxdb-exporter:1.0",
+            "rocks.inspectit:opencensus-influxdb-exporter:1.1",
 			"org.influxdb:influxdb-java:2.15"
 	)
 	compileOnly "org.projectlombok:lombok:1.18.4"

components/inspectit-ocelot-eum-server/src/main/java/rocks/inspectit/oce/eum/server/exporters/InfluxExporterService.java

@@ -50,16 +50,19 @@ private boolean shouldEnable() {
     private void doEnable() {
         InfluxExporterSettings influx = config.getExporters().getMetrics().getInflux();
         if (shouldEnable()) {
-            log.info("Starting InfluxDB Exporter to '{}:{}' on '{}'", influx.getDatabase(), influx.getRetentionPolicy(), influx.getUrl());
+            log.info("Starting InfluxDB Exporter to '{}:{}' on '{}'", influx.getDatabase(), influx.getRetentionPolicy(), influx
+                    .getUrl());
             activeExporter = InfluxExporter.builder()
                     .url(influx.getUrl())
                     .database(influx.getDatabase())
                     .retention(influx.getRetentionPolicy())
                     .user(influx.getUser())
                     .password(influx.getPassword())
                     .createDatabase(influx.isCreateDatabase())
+                    .exportDifference(influx.isCountersAsDifferences())
                     .build();
-            exporterTask = executor.scheduleAtFixedRate(activeExporter::export, 0, influx.getExportInterval().toMillis(), TimeUnit.MILLISECONDS);
+            exporterTask = executor.scheduleAtFixedRate(activeExporter::export, 0, influx.getExportInterval()
+                    .toMillis(), TimeUnit.MILLISECONDS);
         }
     }
 

components/inspectit-ocelot-eum-server/src/main/resources/application.yml

@@ -114,6 +114,11 @@ inspectit-eum-server:
         retention-policy: "autogen"
         # If true, the specified database will be created with the autogen retention policy
         create-database: true
+        # If disabled, the raw values of each counter will be written to the InfluxDB on each export.
+        # When enabled, only the change of the counter in comparison to the previous export will be written.
+        # This difference will only be written if the counter has changed (=the difference is non-zero).
+        # This can greatly reduce the total data written to influx and makes writing queries easier.
+        counters-as-differences: true
 
   # settings for the EUM server's self-monitoring
   self-monitoring:

inspectit-ocelot-config/src/main/java/rocks/inspectit/ocelot/config/model/exporters/metrics/InfluxExporterSettings.java

@@ -54,4 +54,12 @@ public class InfluxExporterSettings {
      * If enabled, the Influx Exporter creates the specified database upon initial connection.
      */
     private boolean createDatabase;
+
+    /**
+     * If disabled, the raw values of each counter will be written to the InfluxDB on each export.
+     * When enabled, only the change of the counter in comparison to the previous export will be written.
+     * This difference will only be exposed if the counter has changed (=the difference is non-zero).
+     * This can greatly reduce the total data written to influx and makes writing queries easier.
+     */
+    private boolean countersAsDifferences;
 }

inspectit-ocelot-config/src/main/resources/rocks/inspectit/ocelot/config/default/exporters.yml

@@ -40,6 +40,11 @@ inspectit:
         retention-policy: "autogen"
         # If true, the specified database will be created with the autogen retention policy
         create-database: true
+        # If disabled, the raw values of each counter will be written to the InfluxDB on each export.
+        # When enabled, only the change of the counter in comparison to the previous export will be written.
+        # This difference will only be written if the counter has changed (=the difference is non-zero).
+        # This can greatly reduce the total data written to influx and makes writing queries easier.
+        counters-as-differences: true
 
     # settings for trace exporters
     tracing:

inspectit-ocelot-core/build.gradle

@@ -38,7 +38,6 @@ dependencies {
             project(':inspectit-ocelot-config'),
             project(':inspectit-ocelot-sdk'),
 
-            "rocks.inspectit:opencensus-influxdb-exporter:1.0",
 
             // spring releated
             "org.springframework.boot:spring-boot:${springVersion}",
@@ -69,6 +68,7 @@ dependencies {
             "io.opencensus:opencensus-exporter-trace-jaeger:${openCensusVersion}",
             "io.opencensus:opencensus-exporter-metrics-ocagent:${openCensusVersion}",
             "io.opencensus:opencensus-exporter-trace-ocagent:${openCensusVersion}",
+            "rocks.inspectit:opencensus-influxdb-exporter:1.1",
 
             // bytecode manipulation
             "net.bytebuddy:byte-buddy:1.9.2"

inspectit-ocelot-core/src/main/java/rocks/inspectit/ocelot/core/exporter/InfluxExporterService.java

@@ -33,7 +33,6 @@ public class InfluxExporterService extends DynamicallyActivatableService {
      */
     private Future exporterTask;
 
-
     public InfluxExporterService() {
         super("exporters.metrics.influx", "metrics.enabled");
     }
@@ -51,16 +50,19 @@ protected boolean checkEnabledForConfig(InspectitConfig conf) {
     @Override
     protected boolean doEnable(InspectitConfig configuration) {
         InfluxExporterSettings influx = configuration.getExporters().getMetrics().getInflux();
-        log.info("Starting InfluxDB Exporter to '{}:{}' on '{}'", influx.getDatabase(), influx.getRetentionPolicy(), influx.getUrl());
+        log.info("Starting InfluxDB Exporter to '{}:{}' on '{}'", influx.getDatabase(), influx.getRetentionPolicy(), influx
+                .getUrl());
         activeExporter = InfluxExporter.builder()
                 .url(influx.getUrl())
                 .database(influx.getDatabase())
                 .retention(influx.getRetentionPolicy())
                 .user(influx.getUser())
                 .password(influx.getPassword())
                 .createDatabase(influx.isCreateDatabase())
+                .exportDifference(influx.isCountersAsDifferences())
                 .build();
-        exporterTask = executor.scheduleAtFixedRate(activeExporter::export, 0, influx.getExportInterval().toMillis(), TimeUnit.MILLISECONDS);
+        exporterTask = executor.scheduleAtFixedRate(activeExporter::export, 0, influx.getExportInterval()
+                .toMillis(), TimeUnit.MILLISECONDS);
         return true;
     }
 

inspectit-ocelot-demo/configuration-server/files/all/general.yml

@@ -4,6 +4,6 @@ inspectit:
   metrics:
     # root setting for the polling frequency of all metrics
     # when a metrics has no frequency defined separately, this frequency will be used
-    frequency: 1s
+    frequency: 5s
   logging:
     debug: false # set to true to see more details in the agents logs

inspectit-ocelot-demo/grafana/marketplace/influx/http.json

@@ -58,7 +58,7 @@
   "gnetId": null,
   "graphTooltip": 0,
   "id": null,
-  "iteration": 1580728981189,
+  "iteration": 1587549556162,
   "links": [],
   "panels": [
     {
@@ -179,7 +179,7 @@
           "pattern": "Time",
           "thresholds": [],
           "type": "hidden",
-          "unit": "short"
+          "unit": "none"
         }
       ],
       "targets": [
@@ -201,7 +201,7 @@
           "hide": false,
           "orderByTime": "ASC",
           "policy": "default",
-          "query": "SELECT * FROM (\n SELECT  SUM(total) as za_requests, SUM(times) / SUM(total) AS zb_response_time FROM (\n      SELECT spread(\"count\") AS total, spread(\"sum\") AS times FROM \"http_in_responsetime\" \n      WHERE $timeFilter AND service =~ /^$service$/ AND application =~ /^$application$/ AND http_path =~ /^$path$/  GROUP BY *) GROUP BY http_path fill(0)\n) WHERE za_requests > 0",
+          "query": "SELECT  SUM(total) as za_requests, SUM(times) / SUM(total) AS zb_response_time FROM (\n  SELECT SUM(\"count\") AS total, SUM(\"sum\") AS times FROM \"http_in_responsetime\" \n  WHERE $timeFilter AND service =~ /^$service$/ AND application =~ /^$application$/ AND http_path =~ /^$path$/  GROUP BY http_path\n) GROUP BY http_path",
           "rawQuery": true,
           "refId": "A",
           "resultFormat": "table",
@@ -236,10 +236,9 @@
               "type": "fill"
             }
           ],
-          "hide": false,
           "orderByTime": "ASC",
           "policy": "default",
-          "query": "SELECT zc_error_rate FROM (\nSELECT  SUM(errors) / SUM(total) AS zc_error_rate, SUM(total) as total FROM (\n      SELECT spread(\"count\") AS total, spread(\"errorcount\")AS errors FROM \"http_in_error\" \n      WHERE $timeFilter AND service =~ /^$service$/ AND application =~ /^$application$/ AND http_path =~ /^$path$/  GROUP BY *) \nGROUP BY http_path fill(0)\n) WHERE total > 0 GROUP BY http_path fill(0)",
+          "query": "SELECT  SUM(errors) / SUM(total) AS zc_error_rate FROM (\n  SELECT SUM(\"count\") AS total, SUM(\"errorcount\") AS errors FROM \"http_in_error\" \n  WHERE $timeFilter AND service =~ /^$service$/ AND application =~ /^$application$/ AND http_path =~ /^$path$/  GROUP BY http_path fill(0)\n)\nGROUP BY http_path",
           "rawQuery": true,
           "refId": "B",
           "resultFormat": "table",
@@ -279,7 +278,7 @@
         "y": 1
       },
       "id": 23,
-      "interval": "5s",
+      "interval": "15s",
       "legend": {
         "avg": false,
         "current": false,
@@ -325,7 +324,7 @@
           "legendFormat": "{{http_path}}",
           "orderByTime": "ASC",
           "policy": "default",
-          "query": "SELECT  SUM(times) / SUM(total) AS value FROM (\n      SELECT difference(\"count\") AS total, difference(\"sum\") AS times FROM \"http_in_responsetime\" \n      WHERE $timeFilter AND service =~ /^$service$/  AND application=~ /^$application$/  AND http_path=~ /^$path$/ GROUP BY  *) \nWHERE total > 0\nGROUP BY time($__interval), http_path fill(linear)",
+          "query": "SELECT  SUM(sum) / SUM(count)\nFROM \"http_in_responsetime\" \nWHERE $timeFilter AND service =~ /^$service$/  AND application=~ /^$application$/  AND http_path=~ /^$path$/\nGROUP BY time($__interval), http_path fill(none)",
           "rawQuery": true,
           "refId": "A",
           "resultFormat": "time_series",
@@ -415,7 +414,7 @@
       },
       "hideTimeOverride": true,
       "id": 6,
-      "interval": "1s",
+      "interval": "15s",
       "links": [],
       "mappingType": 1,
       "mappingTypes": [
@@ -471,7 +470,7 @@
           "intervalFactor": 1,
           "orderByTime": "ASC",
           "policy": "default",
-          "query": "SELECT  SUM(total)/$__interval_ms*1000 AS value FROM (\n      SELECT difference(\"count\") AS total FROM \"http_in_responsetime\" \n      WHERE $timeFilter AND service =~ /^$service$/  AND application=~ /^$application$/  AND http_path=~ /^$path$/ GROUP BY *) \nGROUP BY time($__interval) fill(linear)",
+          "query": "SELECT  SUM(count)/$__interval_ms*1000 AS value\nFROM \"http_in_responsetime\" \nWHERE $timeFilter AND service =~ /^$service$/  AND application=~ /^$application$/  AND http_path=~ /^$path$/ \nGROUP BY time($__interval) fill(0)",
           "rawQuery": true,
           "refId": "A",
           "resultFormat": "time_series",
@@ -592,7 +591,7 @@
           "intervalFactor": 1,
           "orderByTime": "ASC",
           "policy": "default",
-          "query": "SELECT  SUM(errors)/SUM(total) AS value FROM (\n      SELECT difference(\"count\") AS total, difference(\"errorcount\") AS errors FROM \"http_in_error\" \n      WHERE $timeFilter AND service =~ /^$service/  AND application=~ /^$application$/  AND http_path=~ /^$path$/ GROUP BY * ) \nGROUP BY time($__interval) fill(0)",
+          "query": "SELECT  SUM(errors) / SUM(total) AS value FROM (\n  SELECT SUM(\"count\") AS total, SUM(\"errorcount\") AS errors FROM \"http_in_error\" \n  WHERE $timeFilter AND service =~ /^$service$/ AND application =~ /^$application$/ AND http_path =~ /^$path$/  GROUP BY time($__interval) fill(0)\n)\nWHERE total > 0\nGROUP BY time($__interval)",
           "rawQuery": true,
           "refId": "A",
           "resultFormat": "time_series",
@@ -774,7 +773,7 @@
           "legendFormat": "",
           "orderByTime": "ASC",
           "policy": "default",
-          "query": "SELECT * FROM (\n SELECT  SUM(total) as za_requests, SUM(times) / SUM(total) AS zb_response_time FROM (\n      SELECT spread(\"count\") AS total, spread(\"sum\") AS times FROM \"http_out_responsetime\" \n      WHERE $timeFilter AND service =~ /^$service$/ AND application =~ /^$application$/ AND http_path =~ /^$path$/  GROUP BY *) GROUP BY http_host, http_path fill(0)\n) WHERE za_requests > 0",
+          "query": "SELECT  SUM(total) as za_requests, SUM(times) / SUM(total) AS zb_response_time FROM (\n  SELECT SUM(\"count\") AS total, SUM(\"sum\") AS times FROM \"http_out_responsetime\" \n  WHERE $timeFilter AND service =~ /^$service$/ AND application =~ /^$application$/ AND http_path =~ /^$path$/  GROUP BY http_path,http_host\n) GROUP BY *",
           "rawQuery": true,
           "refId": "A",
           "resultFormat": "table",
@@ -811,7 +810,7 @@
           ],
           "orderByTime": "ASC",
           "policy": "default",
-          "query": "SELECT zc_error_rate FROM (\nSELECT  SUM(errors) / SUM(total) AS zc_error_rate, SUM(total) as total FROM (\n      SELECT spread(\"count\") AS total, spread(\"errorcount\")AS errors FROM \"http_out_error\" \n      WHERE $timeFilter AND service =~ /^$service$/ AND application =~ /^$application$/ AND http_path =~ /^$path$/  GROUP BY *) \nGROUP BY http_host,http_path fill(0)\n) WHERE total > 0 GROUP BY http_host, http_path fill(0)",
+          "query": "SELECT  SUM(errors) / SUM(total) AS zc_error_rate FROM (\n  SELECT SUM(\"count\") AS total, SUM(\"errorcount\") AS errors FROM \"http_out_error\" \n  WHERE $timeFilter AND service =~ /^$service$/ AND application =~ /^$application$/ AND http_path =~ /^$path$/  GROUP BY http_path,http_host fill(0)\n)\nGROUP BY *",
           "rawQuery": true,
           "refId": "B",
           "resultFormat": "table",
@@ -851,6 +850,7 @@
         "y": 6
       },
       "id": 76,
+      "interval": "15s",
       "legend": {
         "avg": false,
         "current": false,
@@ -891,7 +891,7 @@
           ],
           "orderByTime": "ASC",
           "policy": "default",
-          "query": "SELECT  SUM(times) / SUM(total) AS value FROM (\n      SELECT difference(\"count\") AS total, difference(\"sum\") AS times FROM \"http_out_responsetime\" \n      WHERE $timeFilter AND service =~ /^$service$/  AND application=~ /^$application$/  AND http_path=~ /^$path$/ GROUP BY  *) \nWHERE total > 0\nGROUP BY time($__interval), http_path,http_host fill(linear)",
+          "query": "SELECT  SUM(sum) / SUM(count) AS value \nFROM \"http_out_responsetime\" \nWHERE $timeFilter AND service =~ /^$service$/  AND application=~ /^$application$/  AND http_path=~ /^$path$/\nGROUP BY time($__interval), http_path,http_host fill(none)",
           "rawQuery": true,
           "refId": "A",
           "resultFormat": "time_series",
@@ -1037,7 +1037,7 @@
           "intervalFactor": 1,
           "orderByTime": "ASC",
           "policy": "default",
-          "query": "SELECT  SUM(total)/$__interval_ms*1000 AS value FROM (\n      SELECT difference(\"count\") AS total FROM \"http_out_responsetime\" \n      WHERE $timeFilter AND service =~ /^$service$/  AND application=~ /^$application$/  AND http_path=~ /^$path$/ GROUP BY * ) \nGROUP BY time($__interval) fill(linear)",
+          "query": "SELECT  SUM(count)/$__interval_ms*1000 AS value\nFROM \"http_out_responsetime\" \nWHERE $timeFilter AND service =~ /^$service$/  AND application=~ /^$application$/  AND http_path=~ /^$path$/ \nGROUP BY time($__interval) fill(0)",
           "rawQuery": true,
           "refId": "A",
           "resultFormat": "time_series",
@@ -1158,7 +1158,7 @@
           "intervalFactor": 1,
           "orderByTime": "ASC",
           "policy": "default",
-          "query": "SELECT  SUM(errors)/SUM(total) AS value FROM (\n      SELECT difference(\"count\") AS total, difference(\"errorcount\") AS errors FROM \"http_out_error\" \n      WHERE $timeFilter AND service =~ /^$service/  AND application=~ /^$application$/  AND http_path=~ /^$path$/ GROUP BY *) \nGROUP BY time($__interval) fill(0)",
+          "query": "SELECT  SUM(errors) / SUM(total) AS value FROM (\n  SELECT SUM(\"count\") AS total, SUM(\"errorcount\") AS errors FROM \"http_out_error\" \n  WHERE $timeFilter AND service =~ /^$service$/ AND application =~ /^$application$/ AND http_path =~ /^$path$/  GROUP BY time($__interval) fill(0)\n)\nWHERE total > 0\nGROUP BY time($__interval)",
           "rawQuery": true,
           "refId": "A",
           "resultFormat": "time_series",
@@ -1270,7 +1270,7 @@
     ]
   },
   "time": {
-    "from": "now-15m",
+    "from": "now-5m",
     "to": "now"
   },
   "timepicker": {
@@ -1301,5 +1301,5 @@
   "timezone": "",
   "title": "HTTP Overview",
   "uid": "523wqxqmz",
-  "version": 2
+  "version": 5
 }

inspectit-ocelot-demo/grafana/marketplace/influx/jvm.json

@@ -58,7 +58,7 @@
   "gnetId": null,
   "graphTooltip": 0,
   "id": null,
-  "iteration": 1580138904831,
+  "iteration": 1587549648580,
   "links": [],
   "panels": [
     {
@@ -890,7 +890,7 @@
           "measurement": "jvm_gc_memory_allocated",
           "orderByTime": "ASC",
           "policy": "default",
-          "query": "SELECT SUM(val) FROM (SELECT NON_NEGATIVE_DIFFERENCE(\"counter\") AS val FROM \"jvm_gc_memory_allocated\" WHERE (\"service\" =~ /^$service$/) AND $timeFilter GROUP BY *) GROUP BY time($__interval)",
+          "query": "SELECT SUM(\"counter\") FROM \"jvm_gc_memory_allocated\" WHERE (\"service\" =~ /^$service$/) AND $timeFilter GROUP BY time($__interval)",
           "rawQuery": true,
           "refId": "D",
           "resultFormat": "time_series",
@@ -934,7 +934,7 @@
           "measurement": "jvm_gc_memory_promoted",
           "orderByTime": "ASC",
           "policy": "default",
-          "query": "SELECT SUM(val) FROM (SELECT NON_NEGATIVE_DIFFERENCE(\"counter\") AS val FROM \"jvm_gc_memory_promoted\" WHERE (\"service\" =~ /^$service$/) AND $timeFilter GROUP BY *) GROUP BY time($__interval)",
+          "query": "SELECT SUM(\"counter\")  FROM \"jvm_gc_memory_promoted\" WHERE (\"service\" =~ /^$service$/) AND $timeFilter GROUP BY time($__interval)",
           "rawQuery": true,
           "refId": "F",
           "resultFormat": "time_series",
@@ -1648,7 +1648,7 @@
           "measurement": "jvm_gc_pause",
           "orderByTime": "ASC",
           "policy": "default",
-          "query": "SELECT SUM(times) FROM (SELECT non_negative_difference(\"counter\") AS times FROM \"jvm_gc_pause\" WHERE (\"service\" =~ /^$service$/ AND \"action\" = 'end of minor GC') AND $timeFilter GROUP BY *) WHERE times > 0 GROUP BY time($__interval) fill(none)",
+          "query": "SELECT SUM(\"counter\") FROM \"jvm_gc_pause\" WHERE (\"service\" =~ /^$service$/ AND \"action\" = 'end of minor GC') AND $timeFilter GROUP BY time($__interval) fill(none)",
           "rawQuery": true,
           "refId": "C",
           "resultFormat": "time_series",
@@ -1698,7 +1698,7 @@
           "measurement": "jvm_gc_pause",
           "orderByTime": "ASC",
           "policy": "default",
-          "query": "SELECT SUM(times) FROM (SELECT non_negative_difference(\"counter\") AS times FROM \"jvm_gc_pause\" WHERE (\"service\" =~ /^$service$/ AND \"action\" = 'end of major GC') AND $timeFilter GROUP BY *) WHERE times > 0 GROUP BY time($__interval) fill(none)",
+          "query": "SELECT SUM(\"counter\") FROM \"jvm_gc_pause\" WHERE (\"service\" =~ /^$service$/ AND \"action\" = 'end of major GC') AND $timeFilter GROUP BY time($__interval) fill(none)",
           "rawQuery": true,
           "refId": "A",
           "resultFormat": "time_series",
@@ -1959,5 +1959,5 @@
   "timezone": "",
   "title": "JVM Metrics",
   "uid": "jaEIHn8ik",
-  "version": 1
+  "version": 2
 }