SOFAStack
S calable O PEN F inancial A rchitecture Stack ants gold dress self-developed financial level distributed architecture, including the various components needed to build a financial level cloud native architecture, best practices forged on the anvil financial scene.
SOFALookout ants gold dress developed lightweight middleware measure and monitor the problem of solving a system of open source service within SOFAStack system. This paper to introduce the next major characteristics SOFALookout server provided and use.
SOFALookout: Githubkcom/sofastchk/skkk
Foreword
Container, K8S, micro service, Mesh and Serverless these new technologies is a fundamental change in the direction of the software we run the way. We build more distributed system of the other due to the life cycle of container, the system is more short, become perishable. In response to these changes, SOFALookout want to provide a lightweight solution. Before SOFALookout has the ability to open source client. Today, SOFALookout server Metrics section of the code finally officially open it! This article give you SOFALookout server-side main features and how to use the next.
What is SOFALookout
SOFALookout is lightweight middleware services measure and monitor the problem of ants gold dress solve a system of open source. It provides services including: Metrics buried point, the collection, processing, storage and inquiries. The open source project consists of two separate parts, namely the client and server-side service.
SOFALookout goal is to create a lightweight platform Observability real-time tools to help users solve the problem of monitoring and analysis infrastructure, applications and services, etc. SOFALookout (now open source part) is a target system to measure and monitor projects using the multi-dimensional metrics. SOFALookout multi-dimensional metrics reference Metrics2.0 standards.
SOFALookout: Githubkcom/sofastchk/skkk
SOFALookout installation documentation: www.sofastack.tech/sofa-lookou...
Key features SOFALookout server:
- Community Main Metrics adapted to write protocol data source (for example: Prometheus , Metricbeat the like);
- Support the expansion of data storage, temporarily open source version supported by default elasticsearch , relevant and transparent and automated operation and maintenance operations;
- Prometheus follow standard API query and support PromQL , and make appropriate improvements;
- It comes with the console data queries, and supports Grafana data visualization;
- Simple to use, supports a single server process running the entire module.
As SOFALookout (metrics) open source server-side code, data has been processed metrics form a closed loop. We will further follow-up of open source service capability and Event Trace related, so stay tuned.
SOFALookout project structure
Each server-side code consists of two parts: Gateway Server module and module. As shown below (module shows a schematic configuration SOFALookout source projects)
├── boot
├── client
├── gateway
└── server复制代码The combined effect of boot module project is to facilitate the integration and operation of the server module, you can run both Gateway and Server services alone, you can make use of SOFAArk complete (Gateway and Server) The All in One runs as a single process.
SOFALookout working mechanism
The following figure shows how SOFALookout complete data acquisition from the metrics, reported and stored in the complete flow path to the final display.
Currently SOFALookout support flexible metrics data storage selection. But we do open source version supports only the Elasticsearch as storage solutions (follow-up may continue to support Cassandra, InfluxDB ...), other storage fit we hope that more students can participate and build support. Priority Support Elasticsearch is because we take into account the ELK solutions have been widely used in the industry, especially the log data.
For out of the box, taking into account the use unfamiliar Elasticsearch classmates, SOFALookout has built automated operation and maintenance tools on metrics data storage, they can build their own exempt Index, and routine maintenance ES Index of trouble, more details follow explain alone.
The new open source modules
A, SOFALookout Gateway Module
SOFALookout Gateway lightweight data pipe that provides extensive support access protocols, including its own SDK (SOFALookout Client) report protocol data protocol also supports the Prometheus (push mode and a pull mode), Metricbeat Protocol (version 6), OpenTSDB written agreement. For each data source corresponding to a concept of Importer.
SOFALookout Gateway for remote (push mode) reported providing local hard disk buffering support. Gateway Pipeline overall design is in the form of data processed around, including pre- and post-data filter developers to facilitate data processing. Further Gateway can support custom Exporter, offers default Elasticsearch Exporter, Standard Exporter (for relaying data between Gateway), a developer can customize or other storage Kafka variety Exporter.
Two, SOFALookout Server Module
SOFALookout Server is compatible with and enhances the RESTful API data and metadata Prometheus queries. We also corresponds PromQL basically compatible and enhanced (not including Alert related syntax), promQL related to parsing logic SOFALookout is transplanted from from Prometheus, made some optimization and improvement, thanks to Prometheus open source so easy to use and powerful golang QL version of realization.
In order to facilitate enabling developers to easily make data exploration and testing, we also provide support for its own Web-UI is able to meet the basic functions.
We recommend you use Grafana data show. Grafana integrated SOFALookout is very simple, just select the protocol to Prometheus as a data source (SOFALookout default query port is: 9090). The following figure shows Grafana new data source configuration:
Recent plans
The figure is the recent Roadmap:
Very welcome more students to participate SOFALookout build, especially to support more Metrics repository.