Prometheus

The tech-radar item "Prometheus" covers several things in Allegro: standard for describing metrics (largely compatible with OpenMetrics) based on the use of names and labels requests_count{dc="dc4",scid="123",endpoint="/login"}) standard for collecting metrics (scraping metrics from the service endpoint - usually /status/prometheus), or sending them ( so-called pushgateway) standard for querying metrics (promQL language) We use 5 applications in monitoring team, that are compliant with this standard: prometheus (currently mainly as a scraper and a tool for generating recording_rules, which are intermediary "metrics from metrics") vmagent (as a tool for controlling the flow of the metrics stream between scrapers and storages - in the future also as a scraper) VictoriaMetrics (as our main metrics storage - a TSDB type database, and as the main tool for responding to promQL queries) promxy (proxy of promQL queries, capable of merging responses from multiple sources - e.g. clusters with different retention) cortex (as an alternative metrics storage - currently in the testing phase at Allegro).