Interconnecting TSDBs: Prometheus and TDengine

Prometheus is an open-source monitoring and alerting toolkit widely adopted for its powerful time-series data model, flexible query language (PromQL), and efficient pull-based architecture. Prometheus excels at collecting metrics from a wide range of systems and applications, making it a go-to solution for IT infrastructure and application monitoring.

TDengine is a high-performance time-series database purpose-built for industrial and real-time analytics scenarios. It features a compact storage engine, native stream processing, and SQL compatibility. With built-in connectors and low resource consumption, TDengine provides a lightweight yet powerful backend for managing massive volumes of time-stamped data efficiently.

This article demonstrates how to interconnect these systems in two ways:

1. remote_write / remote_read from Prometheus → TDengine
2. TDengine’s /metrics endpoint scraped by Prometheus

Prerequisites

Make sure the following are in place:

• A running TDengine cluster
• taosAdapter installed and running
• Prometheus installed and running

Remote Read and Write Between Prometheus and TDengine

1. Create the prometheus_data database in TDengine:

   CREATE DATABASE prometheus_data;

2. Add the following configuration items to prometheus.yml:

   remote_write:
     - url: "http://<taosadapter-url>:6041/prometheus/v1/remote_write/prometheus_data"
       basic_auth:
         username: <tdengine-user>
         password: <tdengine-password>
   
   remote_read:
     - url: "http://<taosadapter-url>:6041/prometheus/v1/remote_read/prometheus_data"
       basic_auth:
         username: <tdengine-user>
         password: <tdengine-password>
       remote_timeout: 10s
       read_recent: true

   Modify the URL, username, and password to match your TDengine deployment.

3. Restart Prometheus:

   sudo systemctl restart prometheus

4. In TDengine, view the metrics in the prometheus_data database:

   USE prometheus_data;
   SHOW STABLES;
   SELECT * FROM metrics LIMIT 10;

You should see frequent samples like go_gc_duration_seconds confirming that remote_write is working.

Monitoring TDengine with Prometheus

taosKeeper exports TDengine metrics in Prometheus format at the /metrics endpoint so that you can monitor your TDengine cluster in Prometheus.

1. Ensure taosKeeper is running:

   sudo systemctl status taoskeeper

2. Test the metrics endpoint:

   curl http://<taoskeeper_host>:6041/metrics

   You’ll see Prometheus‐style output, for example:

   taos_connections 5
   taos_query_latency_seconds{quantile="0.5"} 0.002

3. In prometheus.yml, add a scrape config for taosKeeper:

   scrape_configs:
     - job_name: "tdengine"
       static_configs:
         - targets: ["<taoskeeper_host>:6041"]
   

4. Restart Prometheus:

   sudo systemctl restart prometheus

You can now see TDengine metrics in Prometheus.