diff --git a/.gitignore b/.gitignore
index 0621410a592efcee9a9bc7e4a84058aab67bd735..e484e18c253b9bd16ef338e92f95bf5eb0bda35f 100644
--- a/.gitignore
+++ b/.gitignore
@@ -4,7 +4,7 @@ vidjil/
__pycache__
.env
-volumes/
+volumes*/
volumes_dev/
certs/*.pem
diff --git a/.gitlab-ci.yml b/.gitlab-ci.yml
new file mode 100644
index 0000000000000000000000000000000000000000..9f65577ed6516d063363c074ea2ae760051ceab5
--- /dev/null
+++ b/.gitlab-ci.yml
@@ -0,0 +1,30 @@
+# Metrics
+
+stages :
+ - publish_release
+
+
+deploy_docker:
+ stage: publish_release
+ image: docker:latest
+ services:
+ - name: docker:dind
+ alias: docker
+ command: ["--tls=false"]
+ services:
+ - docker:dind
+ script:
+ - git submodule init
+ - git submodule update --remote
+ - export CUR_DATE=`date +%Y-%m-%d`
+ - export SHA=`echo $CI_COMMIT_SHA | awk '{ print substr($0, 0, 8) }'`
+ - export TAG="vidjil/metrics:$CUR_DATE-$SHA"
+ - docker login -u "$CI_DOCKER_USER" -p "$CI_DOCKER_PASSWORD" $CI_DOCKER
+ - docker build --no-cache -t $TAG -f metrics/Dockerfile ./metrics/
+ - docker push $TAG
+ - docker tag $TAG "vidjil/metrics:latest"
+ - docker push "vidjil/metrics:latest"
+ - docker rmi $TAG "vidjil/metrics:latest"
+ # when: manual
+ tags:
+ - cidocker
diff --git a/docker-compose-node.yml b/docker-compose-node.yml
new file mode 100644
index 0000000000000000000000000000000000000000..23b0c7096617f2cfbf6820c63227ef18aa693b94
--- /dev/null
+++ b/docker-compose-node.yml
@@ -0,0 +1,12 @@
+version: '3.3'
+
+services:
+
+ ## This service allow to get cpu/memory/... information.
+ # It should be launch on target server and not in this project
+ node_exporter:
+ image: prom/node-exporter:v0.18.1
+ container_name: monitoring_node_exporter
+ hostname: monitoring_node_exporter
+ ports:
+ - "9100:9100"
diff --git a/docker-compose.dev.yml b/docker-compose.dev.yml
index 78c113ea950ebec993b7ded78b0586bb23c567fd..5d648a37a8bcf3cc1faf8d11e279a517bf45a860 100644
--- a/docker-compose.dev.yml
+++ b/docker-compose.dev.yml
@@ -1,6 +1,6 @@
version: '3.3'
services:
- metrics_vidjil:
+ vidjil:
volumes:
- ./metrics/entrypoints:/entrypoints
- ./metrics/vidjil/tools/api_vidjil.py:/app/vidjil/tools/api_vidjil.py
diff --git a/docker-compose.yml b/docker-compose.yml
index 0c1c3d023ade04049f1a25e5618ac0276ed5e913..5bccfcca669b313bb2c9eee6ef43b12cc7a3fc46 100644
--- a/docker-compose.yml
+++ b/docker-compose.yml
@@ -5,7 +5,11 @@ services:
prometheus:
image: prom/prometheus:latest
container_name: monitoring_prometheus
- user: "1000:1000"
+ user: root
+ # Commented ports to be only avaialble fron metrics network and not exposed
+ # If needed for dev, you can expose this port but data are fully exposed
+ # ports:
+ # - "9090:9090"
deploy:
replicas: 1
env_file:
@@ -18,7 +22,10 @@ services:
- '--config.file=/etc/prometheus/prometheus.yml'
- '--storage.tsdb.path=/prometheus'
- '--storage.tsdb.retention.time=365d'
- network_mode: "host"
+ # network_mode: "host"
+ networks:
+ - metrics
+ hostname: monitoring_prometheus
# alertmanager:
# image: prom/alertmanager:v0.18.0
@@ -35,7 +42,9 @@ services:
grafana:
image: grafana/grafana:latest
container_name: monitoring_grafana
- user: "1000:1000"
+ user: root
+ ports:
+ - "3000:3000"
environment:
- GF_PATHS_CONFIG=/etc/grafana/grafana.ini
env_file:
@@ -50,7 +59,10 @@ services:
- ./grafana/dashboards.yml:/etc/grafana/provisioning/dashboards/dashboards.yml
- ./grafana/grafana.ini:/etc/grafana/grafana.ini
- ./grafana/dashboards:/var/lib/grafana/dashboards
- network_mode: "host"
+ # network_mode: "host"
+ networks:
+ - metrics
+ hostname: monitoring_grafana
# blackbox:
# image: prom/blackbox-exporter:v0.14.0
@@ -61,16 +73,14 @@ services:
# - ./blackbox:/config
# network_mode: "host"
- # node_exporter:
- # image: prom/node-exporter:v0.18.1
- # container_name: monitoring_node_exporter
- # deploy:
- # replicas: 1
- # network_mode: "host"
- metrics_vidjil:
- image: vidjil/metrics-app:latest
- container_name: monitoring_vidjil_instance
+ vidjil:
+ image: vidjil/metrics:latest
+ container_name: monitoring_vidjil
+ # Commented ports to be only avaialble fron metrics network and not exposed
+ # If needed for dev, you can expose this port but data are fully exposed
+ # ports:
+ # - "5000:5000"
env_file:
# cp .env-sample as .env file and modify it with user/password of metrics declared user on server
# Don't use an admin account. See XXX page to create metrics user if not already done
@@ -85,9 +95,16 @@ services:
- ./certs:/app/certs
- ./metrics/conf.py:/app/conf.py
command: bash entrypoints/flask-entrypoint.sh
- network_mode: "host"
+ # network_mode: "host"
+ networks:
+ - metrics
+ hostname: monitoring_vidjil
build:
./metrics/
+
# volumes:
# prometheus-data:
- # grafana-data:
\ No newline at end of file
+ # grafana-data:
+
+networks:
+ metrics:
\ No newline at end of file
diff --git a/grafana/datasource.yml b/grafana/datasource.yml
index a0ea46ad23f6d38507ad264d12edd22b83f76723..b67f29a28bee2340d31943da35ffdc99abb23043 100644
--- a/grafana/datasource.yml
+++ b/grafana/datasource.yml
@@ -4,7 +4,7 @@ datasources:
- name: Prometheus
type: prometheus
access: proxy
- url: http://localhost:9090
+ url: http://monitoring_prometheus:9090
# - name: Prometheus-perso2
# type: prometheus-perso2
# access: proxy
diff --git a/metrics/Makefile b/metrics/Makefile
new file mode 100644
index 0000000000000000000000000000000000000000..cd6143b6f0ef4508d64515c0faa5e08119831483
--- /dev/null
+++ b/metrics/Makefile
@@ -0,0 +1,14 @@
+init_submodule:
+ git submodule init
+ git submodule update --remote
+
+build_metrics:
+ docker compose build vidjil
+
+init:
+ python3 -m pip install -U pip
+ python3 -m pip install --no-cache -r requirements.txt
+
+unit:
+ #python3 -m pytest --cov=./ --cov-report html:cov.html -v -s .
+ python3 -m unittest test_main.py
\ No newline at end of file
diff --git a/metrics/conf.py b/metrics/conf.sample.py
similarity index 74%
rename from metrics/conf.py
rename to metrics/conf.sample.py
index ab099b7e666bd666359f09b0afc49bcb51b72536..522d66983ba5024f9a88115be6f7c89f3d5f8a87 100644
--- a/metrics/conf.py
+++ b/metrics/conf.sample.py
@@ -10,9 +10,6 @@ vidjil_servers = [
"certificat": "./certs/cert_localhost.pem",
"user":"metrics@vidjil.org",
"password": "foobartest"
- # }, {
- # "url": "https://app.vidjil.org/vidjil/",
- # "certificat": "./certs/cert_app.vidjil.org.pem"
}
]
diff --git a/metrics/main.py b/metrics/main.py
index 600bf4460dbe48dc8cb9d855d943d9ec1e74f534..e9c9c9b0fb448fdadd5929b44da40edfdb0b053e 100644
--- a/metrics/main.py
+++ b/metrics/main.py
@@ -18,8 +18,6 @@ def get_env_list():
@app.route('/metrics_test')
def metrics_test():
content = ""
- # user = os.getenv('METRICS_USER_EMAIL')
- # password = os.getenv('METRICS_USER_PASSWORD')
dev_mode = os.getenv('DEV_MODE')
print( "====== /metrics_test")
@@ -44,7 +42,7 @@ def unavailableServer(url) -> str:
return f"\
# HELP vidjil_api_request_unavailable_server {'A server is unreachable'}\n\
# TYPE vidjil_api_request_unavailable_server {'gauge'}\n\
-vidjil_api_request_unavailable_server {{server={cleanUrlServer(url)}}}"
+vidjil_api_request_unavailable_server {{server='{cleanUrlServer(url)}'}}\n"
@app.route('/metrics')
@@ -65,7 +63,7 @@ def metrics():
vidjil = Vidjil(url_server, ssl=certificat)
vidjil.login(user, password)
- metrics = vidjil.metrics()
+ metrics = vidjil.metrics(metrics_list="all")
# print(f"{metrics=}")
formated_response += formated_metrics(metrics,cleanUrlServer(url_server))
except Exception as e:
@@ -112,6 +110,9 @@ def formated_metrics(metrics, url_server):
'set_generics_by_group': {"type": "gauge", "description": "set_generics_by_group.", "function": convert_set_generics_by_group},
'set_runs_by_group': {"type": "gauge", "description": "set_runs_by_group.", "function": convert_set_runs_by_group},
'set_patients_by_group': {"type": "gauge", "description": "set_patients_by_group.", "function": convert_set_patients_by_group},
+
+ ### Not used
+ # 'request_times': {"type": "gauge", "description": "request_times.", "function": convert_request_times},
}
string = f"# Information : {metrics['message']}\n"
@@ -205,7 +206,6 @@ def convert_login_count(metrics_key, formater, data, url_server):
# HELP vidjil_api_request_login_count {formater['description']}\n\
# TYPE vidjil_api_request_login_count {formater['type']}\n"
for elt in data:
- #Â print(elt)
if elt['auth_user']['email'] == "metrics@vidjil.org":
continue
string += f"\
@@ -225,7 +225,6 @@ def convert_status_analysis(metrics_key, formater, data, url_server):
# HELP vidjil_api_request_status_analysis {formater['description']}\n\
# TYPE vidjil_api_request_status_analysis {formater['type']}\n"
for elt in data:
- print(elt)
string += f"\
vidjil_api_request_status_analysis\
{{status=\"{elt['scheduler_task']['status']}\", task_name=\"{elt['scheduler_task']['task_name']}\", server=\"{url_server}\"}} \
@@ -313,9 +312,11 @@ def convert_sequence_size_by_user(metrics_key, formater, data, url_server):
# HELP vidjil_api_request_sequence_size_by_user {formater['description']}\n\
# TYPE vidjil_api_request_sequence_size_by_user {formater['type']}\n"
for elt in data:
+ # Some old data have none value (no provider filled); needto replace None by 0
+ size = elt['size_file_sum'] if not elt['size_file_sum'] == None else -1
string += f"\
vidjil_api_request_sequence_size_by_user\
-{{user_id=\"{elt['user_id']}\", server=\"{url_server}\"}} {elt['size_file_sum']}\n"
+{{user_id=\"{elt['user_id']}\", server=\"{url_server}\"}} {size}\n"
return string
@@ -365,15 +366,16 @@ vidjil_api_request_set_generics_count\
return string
+# def convert_request_times(metrics_key, formater, data, url_server):
+# return f"=*=*= convert_request_times ==> {data}"
+# string = f"\
+# # HELP vidjil_api_request_request_times {formater['description']}\n\
+# # TYPE vidjil_api_request_request_times {formater['type']}\n"
+# return string
+
def convert_config_analysis_by_users_patients(metrics_key, formater, data, url_server):
- # "status_analysis": [{
- # 'config_id': 2,
- # 'config_name': 'multi+inc+xxx',
- # 'count': 3,
- # 'user_id': 1}, ...]
- # print( "=====\n===== convert_config_analysis_by_users_patients")
- # print(f"{data[:2]=}")
+ # return f"=*=*= convert_config_analysis_by_users_patients ==> {data[:2]}"
string = f"\
# HELP vidjil_api_request_config_analysis_by_users_patients {formater['description']}\n\
# TYPE vidjil_api_request_config_analysis_by_users_patients {formater['type']}\n"
diff --git a/prometheus/prometheus.yml b/prometheus/prometheus.yml
index f61ddff390aefda56ef867b17b57481d26493df4..f5f4605151093d7e80d52dc01df0ca7bafc836f2 100644
--- a/prometheus/prometheus.yml
+++ b/prometheus/prometheus.yml
@@ -1,5 +1,5 @@
global:
- scrape_interval: 15s # Par défaut prometheus va scrapper les cibles toutes les 15 secondes
+ scrape_interval: 90s # Par défaut prometheus va scrapper les cibles toutes les 15 secondes
# Attach these labels to any time series or alerts when communicating with
# external systems (federation, remote storage, Alertmanager).
@@ -21,16 +21,16 @@ global:
# rule_files:
# - "/alertmanager/alert.rules/alerts.rules.yml"
-# la scrape configuration de prometheus, les hosts qu'il va contacter pour chercher les métrique
+
scrape_configs:
- - job_name: 'measure server' # <-- j'ai mis ce titre de job mais le truc c'est de mettre un nom de job explicite
+ - job_name: 'measure server' # Ask for node reporter (cpu, memory, ...) on each targets server at port 9100
static_configs:
- - targets: ['127.0.0.1:9100', '192.168.1.77:9100']
- scrape_interval: 5s
+ - targets: []
+ scrape_interval: 30s
- job_name: 'vidjil'
static_configs:
- - targets: ['127.0.0.1:5000']
+ - targets: ['monitoring_vidjil:5000']
scrape_interval: 90s
diff --git a/readme.md b/readme.md
index 73d3c814ec302909322ca551a9768e1f113fbdcb..aad156affdd49cd2e0f3f757fa3654249d4d434f 100644
--- a/readme.md
+++ b/readme.md
@@ -7,37 +7,80 @@ You can find information on third-party-tools at:
- [Grafana](https://github.com/grafana/grafana)
<!-- - [Alertmanager](https://github.com/prometheus/alertmanager) -->
<!-- - [Blackbox exporter](https://github.com/prometheus/blackbox_exporter) -->
-<!-- - [Node exporter](https://github.com/prometheus/node_exporter) -->
+- [Node exporter](https://github.com/prometheus/node_exporter)
-### Set parameters
+### Set project
+
+#### Vidjil submodules initialisation
+
+This project depend of vidjil component.
+A first step is needed to get vidjil reposirory fetch inside this project.
+
+You need to launch this command to set it up.
+
+```
+make init_submodules
+```
+
+If everything work fine, Vidjil submodule will be downloaded under `metrics/vidjil`.
+
#### Grafana
-J'ai changé le login/pass par défaut, voici les credentials de grafana :
+Default password of grafana server is `admin/admin`.
+At first login, a new password will be asked.
+
+<!-- TODO; filled information to get correct dashboard at starting -->
-- login : `unicorn`
-- password : `UnicornsExists!`
+#### Prometheus
+
+Configuration should be setted in `prometheus/prometheus.yml` file.
+For vidjil metrics, everything should work out of the box.
+However, you need to declare target servers for node exporter in `measure server` job.
#### Vidjil servers
-TODO
-## lancer la stack
+Configuration can be set in a conf file `metrics/conf.py`. A sample if present under `metrics/conf.sample.py`.
+Please copy and rename this file as `conf.py` to be sure to not send your credential on repository.
+You can after that modifiy configuration to set urln, metrics user and password
+
+You can also use a SSL certificate for API as described in [vidjil documentation](https://www.vidjil.org/doc/api/).
+On target server, you need to set up a metrics user if not already done.
+See this [page](https://www.vidjil.org/doc/admin/#server-monitoring) for more information on how to do that.
+
+#### Node reporter
+
+Node reporter isa dedicated service to follow hardware of a server (CPU and memory usage, disk space, ...).
+It should be launch on each target and declared in prometheus as specified below.
+
+You can use for that dedicated `docker-compose-node.yml` file.
+
+
+## Launch services
+
To launch this stack, you will only need to launch a docker-compose file and set some variable on server to follow :
-```console
-docker-compose up -d
+
```
+docker-compose up -d
+```
+
+Note that for the moment, no image is avaialbe for vidjil metrics.
+At first lauch, you should launch `make build_metrics` to build metrics docker images that will be used.
+Note that if you made some dev and want to build again this image, you can add `--build` at `docker compsoe up -d` to build metrics server image at launch.
+
+By default, only grafana server will be serve on port 3000.
+As metrics and prometheus don't allow to manage user and credential,
+we use an internal network allowing a communication between service but not with external world.
+However, you can open network by uncomment `ports` declaration of these services.
Host network will be used. You can find services on **localhost** at these port
- Prometheus : `9090`
- Grafana : `3000`
+- Metrics vidjil: `5000`
+- Node exporter : `9100`
<!-- - Alertmanager : `9093` -->
-<!-- - Node exporter : `9100` -->
-
-
-## ouidou
-Chez ouidou, nous aimons jouer avec les nouvelles technologies et nous sommes toujours curieux de découvrir des projets sympa et de nouvelles personnes. On recrute, donc n'hésite pas à nous contacter [ici](mailto:contact@ouidou.fr) et de visiter notre site [ici](https://ouidou.fr).
<!-- Based on a projet of ouidou found on github -->
<!-- Ce dépôt est en lien avec mon article medium : https://medium.com/ouidou/un-monitoring-complet-en-quelques-minutes-avec-prometheus-33e849e6392e -->
\ No newline at end of file