diff --git a/Dockerfile b/Dockerfile new file mode 100644 index 0000000..527e099 --- /dev/null +++ b/Dockerfile @@ -0,0 +1,6 @@ +FROM python:3.10-slim-buster +WORKDIR /app +COPY tubearchivist-metrics/requirements.txt . +RUN pip3 install -r requirements.txt +COPY . . +CMD ["python3", "-u", "tubearchivist-metrics/main.py"] \ No newline at end of file diff --git a/README.md b/README.md index 5717484..f8b5e61 100644 --- a/README.md +++ b/README.md @@ -3,7 +3,83 @@ Provide Tube Archivist metrics in Prometheus/OpenMetrics format This is an optional service as part of the Tube Archivist stack. -This is WIP +**This is WIP** +--- +  + +## Metrics reported +``` +channel_count = Number of channels +playlist_count = Number of playlists +download_count = Number of downloads +download_queue = Number of pending downloads +subtitle_count = Number of subtitles downloaded for videos + +``` ## Configuration +--- +### Environment variables +``` +ES_URL: The URL to your ElasticSearch server. Defaults to http://archivist-es:9200 +ES_USER: The username for authentication to ElasticSearch. Defaults to elastic +ES_PASSWORD: The password for authentication to ElasticSearch. No default is set. +LISTEN_PORT: The listen port for the metrics server to run on. Defaults to 9934 +POLL_INTERVAL: The interval in seconds for the data to be scraped from ElasticSearch. Defaults to 60 +``` +--- +### Running via Docker Compose (supported) +This metrics service is designed to be ran inside of docker. + +To view the main compose file for TubeArchivist, please see the main repository here: [TA Repo](https://github.com/bbilly1/tubearchivist) + +To add the metrics service in, place this into your compose file and update the environment variables as required. + +``` + archivist-metrics: + image: tubearchivist-metrics:latest + container_name: archivist-metrics + restart: always + environment: + - "ES_USER=elastic" + - "ES_PASSWORD=verysecret" + - "ES_URL=http://archivist-es:9200" + - "LISTEN_PORT=9934" + - "POLL_INTERVAL=60" + ports: + - 9934:9934 +``` +--- + +### Running via Standalone +Should you want to, you can install and run this without docker. + +To do so, clone this repo, install the python libraries with `pip3 install -r requirements.txt` + +then run `python3 tubearchivist-metrics/main.py` + +Environment variables can be passed in via normal syntax for your OS. + +--- + +### Prometheus example config +``` + - job_name: 'tubearchivist-metrics' + metrics_path: / + static_configs: + - targets: + - :9934 +``` +--- +## How are metrics gathered? + +Typically, a prometheus server will poll the HTTP endpoint of the metrics service to obtain its metrics. + +In most scenarios, a service will then retrieve the data for the metric, and then respond to the prometheus http call. However this can be quite harsh on databases and applications, especially when prometheus is polling every 15 seconds. + +To prevent performance issues and unncessecary load on ElasticSearch. We prefetch the metric information from ES every 60 seconds (default). The metric is then updated on the HTTP endpoint after we have retrieved the data and cached for prometheus to scrape. + +This means prometheus can scrape the endpoint every second if it likes, but no database calls to ES will be made until the polling interval is reached. + +If you require more granular polling, you can update the `POLLING_INTERVAL` environment variable \ No newline at end of file diff --git a/deploy.sh b/deploy.sh old mode 100644 new mode 100755 index 5eb6cd4..b505716 --- a/deploy.sh +++ b/deploy.sh @@ -1,3 +1,5 @@ +#!/bin/bash + function validate { if [[ $1 ]]; then @@ -21,10 +23,10 @@ function validate { } -exit 0 - if [[ $1 == "validate" ]]; then validate else echo "valid options are: validate" -fi \ No newline at end of file +fi + +exit 0 \ No newline at end of file diff --git a/tubearchivist-metrics/environment.py b/tubearchivist-metrics/environment.py new file mode 100644 index 0000000..983effa --- /dev/null +++ b/tubearchivist-metrics/environment.py @@ -0,0 +1,32 @@ +""" +Functionality for setting up the environment for the metrics package. +Reads in environment variables for the application to use. +""" +import os + + +class AppConfig: + def __init__(self) -> None: + self.config = self.get_config() + + @staticmethod + def get_config(): + """ + Reads in environment variables for the application to use. + """ + + es_pass = os.environ.get("ES_PASSWORD") + es_user = os.environ.get("ES_USER", default="elastic") + es_url = os.environ.get("ES_URL", default="http://archivist-es:9200") + listen_port = os.environ.get("LISTEN_PORT", default="9934") + poll_interval = os.environ.get("POLL_INTERVAL", default="60") + + application = { + "es_url": es_url, + "es_user": es_user, + "es_pass": es_pass, + "listen_port": listen_port, + "poll_interval": poll_interval, + } + + return application diff --git a/tubearchivist-metrics/esconnect.py b/tubearchivist-metrics/esconnect.py new file mode 100644 index 0000000..07abfe5 --- /dev/null +++ b/tubearchivist-metrics/esconnect.py @@ -0,0 +1,59 @@ +from multiprocessing import AuthenticationError +from elasticsearch import ( + Elasticsearch, + ConnectionError, + ConnectionTimeout, + AuthenticationException, + AuthorizationException, +) + +from environment import AppConfig +from time import sleep + + +class ElasticWrapper: + def handle_err(error): + print("Connection Error: " + str(error)) + print("There was a problem connecting to Elasticsearch") + print( + "Please see the error above. This may be as Elasticsearch is still starting up or a misconfiguration" + ) + print("Sleeping for 10 seconds...") + sleep(10) + + def get_count(index_name): + """ + Returns the number of documents in the index + """ + config = AppConfig().config + es_url = config["es_url"] + es_user = config["es_user"] + es_pass = config["es_pass"] + + es = Elasticsearch( + [es_url], + basic_auth=(es_user, es_pass), + timeout=10, + max_retries=12, + retry_on_timeout=True, + ) + + response = 0 + try: + response = es.count(index=index_name)["count"] + except AuthenticationException as e: + ElasticWrapper.handle_err(e) + except ConnectionError as e: + ElasticWrapper.handle_err(e) + except ConnectionTimeout as e: + ElasticWrapper.handle_err(e) + except AuthenticationError as e: + ElasticWrapper.handle_err(e) + except AuthorizationException as e: + ElasticWrapper.handle_err(e) + except: + print("Unknown error occurred. Check your credentials, url and try again.") + print("Sleeping for 10 seconds...") + sleep(10) + else: + return response diff --git a/tubearchivist-metrics/getmetrics.py b/tubearchivist-metrics/getmetrics.py new file mode 100644 index 0000000..21497b0 --- /dev/null +++ b/tubearchivist-metrics/getmetrics.py @@ -0,0 +1,11 @@ +from esconnect import ElasticWrapper + + +class GetMetrics: + @staticmethod + def count(index_name): + + """Get count of documents from ES""" + result = ElasticWrapper.get_count(index_name) + print("Metric for " + index_name + ": " + str(result)) + return int(result) diff --git a/tubearchivist-metrics/main.py b/tubearchivist-metrics/main.py new file mode 100644 index 0000000..59cc207 --- /dev/null +++ b/tubearchivist-metrics/main.py @@ -0,0 +1,64 @@ +import time +from prometheus_client import start_http_server, Gauge + +from environment import AppConfig +from getmetrics import GetMetrics + +config = AppConfig().config + +# Print configuration on console when starting the application + +print("Configuration is currently set to:") +print("Elasticsearch URL: " + config["es_url"]) +print("Listen Port: " + config["listen_port"]) +print("Poll Interval (seconds): " + config["poll_interval"]) + + +class AppMetrics: + # fmt: off + def __init__(self, poll_interval=int(config["poll_interval"])): + + self.poll_interval = poll_interval + + # Metrics to expose + self.channel_count = Gauge("yta_channel_count", "Number of channels") + self.playlist_count = Gauge("yta_playlist_count", "Number of playlists") + self.download_count = Gauge("yta_download_count", "Number of downloads") + self.download_queue = Gauge("yta_download_queue", "Number of pending downloads") + self.subtitle_count = Gauge("yta_subtitle_count", "Number of subtitles downloaded for videos") + # fmt: on + + def run_metrics_loop(self): + """ + Runs a loop that will update the metrics every second. + """ + while True: + self.retrieve_metrics() + time.sleep(self.poll_interval) + + def retrieve_metrics(self): + """ + Retrieves the metrics from the database and updates the metrics. + """ + print("Obtaining Metrics from Elasticsearch") + self.channel_count.set(GetMetrics.count(index_name="ta_channel")) + self.playlist_count.set(GetMetrics.count(index_name="ta_playlist")) + self.download_queue.set(GetMetrics.count(index_name="ta_download")) + self.download_count.set(GetMetrics.count(index_name="ta_video")) + self.subtitle_count.set(GetMetrics.count(index_name="ta_subtitle")) + + +def main(): + """Main Entry Point""" + listen_port = int(config["listen_port"]) + poll_interval = int(config["poll_interval"]) + + app_metrics = AppMetrics( + poll_interval=poll_interval, + ) + start_http_server(listen_port) + app_metrics.run_metrics_loop() + + +if __name__ == "__main__": + main() diff --git a/tubearchivist-metrics/requirements.txt b/tubearchivist-metrics/requirements.txt new file mode 100644 index 0000000..5902098 --- /dev/null +++ b/tubearchivist-metrics/requirements.txt @@ -0,0 +1,2 @@ +elasticsearch==8.1.2 +prometheus_client==0.14.1