Skip to content

Latest commit

 

History

History
58 lines (32 loc) · 2.03 KB

README.md

File metadata and controls

58 lines (32 loc) · 2.03 KB

EIDA central statistics repository

There are 3 parts of this project :

Aggregator

The aggregator is meant to be deployed on each EIDA node. It is responsible for computing logging statistics with the dataselect usage events as provided by seiscomp.

Centralized statistics web service

This web service is able to receive logging statistics as provided by the aggregator. It stores the logging statistics in a database.

The web service also provide an interface to consult the statistics.

Centralized service management tool

A command line interface to manage the central database system : listing nodes, adding nodes, updating a token.

Test environment

Starting the webserver and the database:

docker-compose up

Deployment

Database

A working postgresql instance is needed, with the hll extension installed. This extension is usually available in the PGDG repository (see for yum or APT). Otherwise, look at the installation instructions of the postgres-hll project

Create a new database and enable the HLL extension :

CREATE ROLE fedstats WITH LOGIN PASSWOD 'xxxxxxxxx';
CREATE DATABASE fedstats OWNER fedstats;
\c fedstats
CREATE EXTENTION hll;

Now you need to deploy the database schema. For this, read the instructions.

Database management tool

eida_statsman from this project will help you manage the nodes and the tokens for statistics ingestion.

To install it, look at it's documentation

Webservice deployment

The recommended way of deploying the webservice is to use the docker image distributed in this project ghcr.io/eida/eida-statistics:main.

To configure and run an instance of the webservice, see it's documentation.

Aggregate and submit statistics

Now, use the aggregator program to build statistics and send them to the webservice instance. Look at the documentation