wip: example dashboard for grafana

README.md

@@ -86,9 +86,10 @@ Example:
 
 ### Layers
 
-We can differentiate the processing of a request in KrakenD in 3 main stages:
+We can differentiate the processing of a request in KrakenD in 3 main stages (each one
+including or wrapping the inner stage):
 
-- `router`: the part that comes before the `Lura`'s framework starts working with
+- `global`: this part that comes before the `Lura`'s framework starts working with
     the request. In the case of [KrakenD CE](https://github.com/krakend/krakend-ce),
     this stage is implemented usin [gin](https://github.com/gin-gonic/gin)
 
@@ -104,7 +105,7 @@ We can differentiate the processing of a request in KrakenD in 3 main stages:
 For each of those layers it can be selected the deatail of metrics and traces
 that we want to report.
 
-#### router
+#### global
 
 At the router level we have 3 main options:
 
@@ -114,7 +115,7 @@ At the router level we have 3 main options:
     traces (so spans from a previous layer are linked to the KrakenD trace).
 
 ```json
-"router": {
+"global": {
     "disable_metrics": false,
     "disable_traces": false,
     "disable_propagation": false
@@ -123,12 +124,12 @@ At the router level we have 3 main options:
 
 ##### Metrics
 
-- `router-response-latency`: histogram of the time it takes to produce the response.
+- `http.server.duration`: histogram of the time it takes to produce the response.
     Attributes:
     - `http.response.status_code`: status code of the produced response
     - `url.path`: the matched endpoint path
     - `krakend.stage`: always with value `router`
-- `router-response-size`: histogram of the size of the body produced for the response.
+- `http.server.response.size`: histogram of the size of the body produced for the response.
     Attributes:
     - `http.response.status_code`: status code of the produced response
     - `url.path`: the matched endpoint path
@@ -151,12 +152,11 @@ At the pipe level we only have 2 options:
 
 ##### Metrics
 
-- `stage-duration`: histogram of the time it takes to produce the response.
+- `krakend.pipe.duration`: histogram of the time it takes to produce the response.
     Attributes:
     - `url.path`: the matched endpoint path that **krakend is serving** (is different
         than in `backend`, krakend stage, when this property is the path
         for the backend we are targetting).
-    - `krakend.stage`: always with value `pipe`
 
 ##### Traces
 
@@ -222,13 +222,11 @@ For both, the `metrics` and `traces` part, we can select the same options:
 
 ##### Metrics
 
-- `stage-duration`: histogram of the time it takes to produce the response. Controlled
+- `krakend.backend.duration`: histogram of the time it takes to produce the response. Controlled
   by the `disable_stage` flag (if set to `true` this metric will not appear).
     Attributes:
     - `url.path`: the matched endpoint path that **krakend is serving** (is different
         than in `backend`, krakend stage, when this property is the path
-        for the backend we are targetting).
-    - `krakend.stage`: always with value `backend`
     - `krakend.endpoint`: this attribute is set to the krakend exposed endpoint that 
         is the "parent" of the backend request.
     - `server.address`: the target host (in case more than one are provided, those
@@ -245,17 +243,18 @@ The following metrics are enabled if `round_trip` is set to true, and share the
         metric, or if we should change that attribute there).
     - `krakend.stage`: always with value `backend-request`
 
+- `http.client.duration`: histogram with the time taken since starting a request, until
+    until having the first byte of the body ready to read.
 
-- `requests-started-count`: number of requests started.
-- `requests-failed-count`: number of requests failed.
-- `requests-canceled-count`: number of canceled request.
-- `requests-timedout-count`: number of timed out requests.
-- `requests-content-length`: counter wit the sum of `Content-Length` header for the 
+- `http.client.request.started.count`: number of requests started.
+- `http.client.request.failed.count`: number of requests [failed](failed).
+- `http.client.request.canceled.count`: number of canceled request.
+- `http.client.request.timedout.count`: number of timed out requests.
+- `http.client.request.content-length`: counter wit the sum of `Content-Length` header for the 
     sent payload for the request.
 
-- `response-latency`: histogram with the time taken until receiving the first byte of the response
-- `response-content-length`: histogram with the size of response bodies as reported in the 
-    `Content-Lenght` header.
+- `http.client.response.content-length`: histogram with the size of response bodies as reported in the 
+    `Content-Length` header.
 
 
 ###### Read Payload metrics
@@ -372,7 +371,7 @@ Putting it all together, here we have an example of a configuration:
         }
     ],
     "layers": {
-        "router": {
+        "global": {
             "disable_metrics": false,
             "disable_traces": false,
             "disable_propagation": false

example/Makefile

@@ -10,6 +10,12 @@ image:
 conf: 
 	cat ./docker_compose/conf.local/prometheus/prometheus.env_tmpl.yml | envsubst > ./docker_compose/conf.local/prometheus/prometheus.yml
 
+docker_up:
+	cd ./docker_compose && docker-compose -p krakend_otel_local -f compose.local.yaml up -d 
+
+docker_down:
+	cd ./docker_compose && docker-compose -p krakend_otel_local -f compose.local.yaml down --remove-orphans -v
+
 run: srv
 	cd ./docker_compose && docker-compose -p krakend_otel_local -f compose.local.yaml up -d 
 	./srv -c ./docker_compose/conf.local/krakend_front/configuration.json  > log_frontend.log &

example/client/k6client/client.js

@@ -0,0 +1,24 @@
+import http from 'k6/http';
+import { sleep } from 'k6';
+
+export const options = {
+  // A number specifying the number of VUs to run concurrently.
+  vus: 10,
+  // A string specifying the total duration of the test run.
+  duration: '15m',
+};
+
+// The function that defines VU logic.
+//
+// See https://grafana.com/docs/k6/latest/examples/get-started-with-k6/ to learn more
+// about authoring k6 scripts.
+//
+export default function() {
+    let address = 'http://192.168.1.12:54444';
+    http.get(address + '/fake/fsf');
+    http.get(address + '/combination/1');
+    http.get(address + '/direct/slow');
+    http.get(address + '/direct/delayed');
+    http.get(address + '/direct/drop');
+    sleep(1);
+}

example/client/make_requests.sh

@@ -0,0 +1,29 @@
+#!/bin/bash
+
+# to generate a new base script:
+#
+# docker run --rm -i -v ${PWD}:/app \
+#     --user $(id -u):$(id -g) \
+#     -w /app grafana/k6 \
+#     new client.js
+
+docker run --rm -i -v ${PWD}:/app \
+    --user $(id -u):$(id -g) \
+    -w /app grafana/k6 \
+    run k6client/client.js
+
+## for i in {1..1}
+## # for i in {1..1000}
+## do
+##     # curl localhost:54444/fake/fsf
+##     # curl localhost:54444/combination/2
+##     curl localhost:54444/combination/1
+##     # curl localhost:54444/direct/slow
+##     sleep 0.1
+##     # curl localhost:54444/direct/delayed
+##     # curl localhost:54444/direct/drop
+## done
+## 
+## # curl localhost:44444/fake/fsf | jq 
+## # echo -e "\n"
+## # curl localhost:44444/combination/23 | jq

example/docker_compose/compose.local.yaml

@@ -8,7 +8,7 @@ services:
     healthcheck:
       test: [CMD, wget, --spicer, -S, http://localhost:3100/ready]
     ports:
-      - "3100:3100"
+      - "53100:3100"
   tempo:
     image: grafana/tempo:latest
     domainname: tempo
@@ -17,24 +17,24 @@ services:
       - "./conf.local/tempo/tempo.yaml:/etc/tempo.yaml"
       - "vol_local_tempo:/tmp/tempo"
     ports:
-      - "14268:14268"   # jaeger ingest
-      - "3200:3200"     # tempo
-      - "9095:9095"     # tempo grpc
-      - "54317:4317"    # otlp grpc
-      - "54318:4318"    # otlp http
-      - "9411:9411"     # zipkin
+      - "53200:3200"     # tempo
+      - "59095:9095"     # tempo grpc
+      - "53317:4317"    # otlp grpc
+      - "53318:4318"    # otlp http
+# - "14268:14268"   # jaeger ingest
+# - "9411:9411"     # zipkin
   prometheus:
     image: prom/prometheus:latest
     domainname: prometheus
     ports:
-      - "59090:9090"
+      - "59590:9090"
     volumes:
       - "./conf.local/prometheus/prometheus.yml:/etc/prometheus/prometheus.yml"
   grafana:
     image: grafana/grafana:latest
     domainname: grafana
     ports:
-      - "3000:3000"
+      - "53000:3000"
     environment:
       GF_SECURITY_ADMIN_USER: krakend
       GF_SECURITY_ADMIN_PASSWORD: krakend
@@ -46,15 +46,19 @@ services:
       - loki
       - tempo
   jaeger:
-    image: jaegertracing/all-in-one:1.52
+    image: jaegertracing/all-in-one:1.54
     domainname: jaeger
     environment:
       COLLECTOR_ZIPKIN_HOST_PORT: ":9411"
     ports:
-      - "5778:5778"   # serve configs 
-      - "16686:16686" # serve frontend UI
+      - "55778:5778"   # serve configs 
+      - "56686:16686"  # serve frontend UI
       - "55317:4317"   # otlp grpc: we remap this to be able to run other envs
       - "55318:4318"   # otlp http: we reamp this to be able to run other envs
+    deploy:
+      resources:
+        limits:
+          memory: 1024M
   fakeapi:
     image: dhontecillas/reqstatsrv:latest
     domainname: fakeapi

example/docker_compose/conf.local/fakeapi/config/example.dockerized.json

@@ -70,12 +70,12 @@
                     "name": "delayer",
                     "config": {
                         "delay_millis_distribution": [
-                            {"key": 0, "val": 0.05},
-                            {"key": 10, "val": 0.2},
+                            {"key": 0, "val": 0.0},
+                            {"key": 10, "val": 0.0},
                             {"key": 50, "val": 0.0},
-                            {"key": 100, "val": 0.5},
-                            {"key": 200, "val": 0.0},
-                            {"key": 700, "val": 0.25}
+                            {"key": 100, "val": 0.25},
+                            {"key": 200, "val": 0.25},
+                            {"key": 700, "val": 0.5}
                         ],
                         "seed": 1
                     }