Monitoring#

Overview#

The monitoring scheme of OpenMLDB is outlined as follows：

  • Use Prometheus to collect monitoring metrics, Grafana to visualize metrics

  • OpenMLDB exporter exposes database-level metrics

  • Each component as a server itself expose component-level metrics

  • Uses node_exporter to expose machine and operating system related metrics

Quick Deployment#

  1. [Optional] Deploy node_exporter on each machine of OpenMLDB. If not deployed, it will not affect the display of Grafana OpenMLDB Dashboard.

  2. [Optional] Deploy an OpenMLDB exporter. If not deployed, it will only result in a few missing data in the Grafana OpenMLDB Dashboard, without affecting the monitoring of read and write operations.

  3. Start Prometheus with the following minimal configuration file. Fill in the corresponding IP addresses, but do not fill in the IP address of TaskManager (which does not support metrics):

global:
  scrape_interval:     15s # By default, scrape targets every 15 seconds.

# A scrape configuration containing exactly one endpoint to scrape:
# Here it's Prometheus itself.
scrape_configs:
  # The job name is added as a label `job=<job_name>` to any timeseries scraped from this config.
  - job_name: openmldb_components
    metrics_path: /brpc_metrics
    static_configs:
      - targets:
        - nameserver_ip
        - tablet_ip
        - tablet_ip
        - apiserver_ip

For complete configuration, refer to openmldb_mixin/prometheus_example.yml.

Command reference: docker run -d -v <config_file>:/etc/prometheus/prometheus.yml -p 9090:9090 -name promethues prom/prometheus

  1. Start Grafana and use the OpenMLDB Dashboard template

Command reference: docker run -d -p 3000:3000 --name=grafana grafana/grafana-oss

Create a Dashboard using the template, Template ID: 17843, URL: https://grafana.com/grafana/dashboards/17843. If it is an empty Dashboard, you can modify the JSON Model in the settings and paste the template content.

  1. To track Deployment execution, you also need to configure the OpenMLDB global variable SET GLOBAL deploy_stats = 'on';.

Install and Run OpenMLDB Exporter#

Introduction#

PyPI PyPI - Python Version

The OpenMLDB exporter is a Prometheus exporter implemented in Python. The core connects the OpenMLDB instance through the database SDK and will query the exposed monitoring indicators through SQL statements. Exporter publish into PyPI. You can install the latest openmldb-exporter through pip. For development instructions, please refer to the code directory README.

Environmental Requirements#

  • Python >= 3.8

  • OpenMLDB >= 0.5.0

Compatibility#

OpenMLDB Exporter version

OpenMLDB supported version

Grafana Dashboard revision

Explaination

>= 0.9.0

>= 0.8.4

>=4

OpenMLDB removed deploy response time in database since 0.8.4

< 0.9.0

>= 0.5.0, < 0.8.4

3

Preparation#

  1. Get OpenMLDB

    You can download precompiled packages from the OpenMLDB release page.

  2. Start OpenMLDB

    See install_deploy How to start OpenMLDB components

    OpenMLDB exporter requires OpenMLDB starts with the server status feature, make sure there is the startup parameter --enable_status_service=true, or --enable_status_service=true appears in startup flag files (usually conf/(tablet|nameserver).flags).

    The startup script bin/start.sh should enable server status by default.

  3. Note: Make sure to select the binding IP addresses of OpenMLDB components OpenMLDB exporter as well as Prometheus and Grafana to ensure that Grafana can access Prometheus, and that Prometheus, OpenMLDB exporter, and OpenMLDB components can access each other.

Deploy the OpenMLDB exporter#

You can run openmdlb-exporter from docker, or install and run directly from PyPI.

Use docker 
docker run ghcr.io/4paradigm/openmldb-exporter \
    --config.zk_root=<openmldb_zk_addr> \
    --config.zk_path=<openmldb_zk_path>
Install and Run from PyPI 
pip install openmldb-exporter

# start
openmldb-exporter \
    --config.zk_root=<openmldb_zk_addr> \
    --config.zk_path=<openmldb_zk_path>

And replace <openmdlb_zk_addr> and <openmldb_zk_path> to correct value. Afterwards, you can check metrics with curl:

curl http://<IP>:8000/metrics

<IP> is docker container IP, or 127.0.0.1 if installing from PyPI.

Example output 
# HELP openmldb_connected_seconds_total duration for a component conncted time in seconds                              
# TYPE openmldb_connected_seconds_total counter                                                                        
openmldb_connected_seconds_total{endpoint="172.17.0.15:9520",role="tablet"} 208834.70900011063                         
openmldb_connected_seconds_total{endpoint="172.17.0.15:9521",role="tablet"} 208834.70700001717                         
openmldb_connected_seconds_total{endpoint="172.17.0.15:9522",role="tablet"} 208834.71399998665                         
openmldb_connected_seconds_total{endpoint="172.17.0.15:9622",role="nameserver"} 208833.70000004768                     
openmldb_connected_seconds_total{endpoint="172.17.0.15:9623",role="nameserver"} 208831.70900011063                     
openmldb_connected_seconds_total{endpoint="172.17.0.15:9624",role="nameserver"} 208829.7230000496                      
# HELP openmldb_connected_seconds_created duration for a component conncted time in seconds                            
# TYPE openmldb_connected_seconds_created gauge                                                                        
openmldb_connected_seconds_created{endpoint="172.17.0.15:9520",role="tablet"} 1.6501813860467942e+09                   
openmldb_connected_seconds_created{endpoint="172.17.0.15:9521",role="tablet"} 1.6501813860495396e+09                   
openmldb_connected_seconds_created{endpoint="172.17.0.15:9522",role="tablet"} 1.650181386050323e+09                    
openmldb_connected_seconds_created{endpoint="172.17.0.15:9622",role="nameserver"} 1.6501813860512116e+09               
openmldb_connected_seconds_created{endpoint="172.17.0.15:9623",role="nameserver"} 1.650181386051238e+09                
openmldb_connected_seconds_created{endpoint="172.17.0.15:9624",role="nameserver"} 1.6501813860512598e+09

Configuration#

You can view the help from:

openmldb-exporter -h

--config.zk_root and --config.zk_path are mandatory.

Available options 
usage: openmldb-exporter [-h] [--log.level LOG.LEVEL] [--web.listen-address WEB.LISTEN_ADDRESS]
                        [--web.telemetry-path WEB.TELEMETRY_PATH] [--config.zk_root CONFIG.ZK_ROOT]
                        [--config.zk_path CONFIG.ZK_PATH] [--config.interval CONFIG.INTERVAL]

OpenMLDB exporter

optional arguments:
 -h, --help            show this help message and exit
 --log.level LOG.LEVEL
                       config log level, default WARN
 --web.listen-address WEB.LISTEN_ADDRESS
                       process listen port, default 8000
 --web.telemetry-path WEB.TELEMETRY_PATH
                       Path under which to expose metrics, default metrics
 --config.zk_root CONFIG.ZK_ROOT
                       endpoint to zookeeper, default 127.0.0.1:6181
 --config.zk_path CONFIG.ZK_PATH
                       root path in zookeeper for OpenMLDB, default /
 --config.interval CONFIG.INTERVAL
                       interval in seconds to pull metrics periodically, default 30.0

Deploy Node Exporter#

node_exporter is an official implementation of Prometheus that exposes system metrics, read their README about setup. To display these metrics in Grafana, use the official Dashboard 1860 provided by Prometheus.

Deploy Prometheus and Grafana#

For installation and deployment of Prometheus and Grafana, please refer to the official documents promtheus get started and [Grafana get started](https://grafana.com/docs/ grafana/latest/getting-started/getting-started-prometheus/). We recommend quick start with docker images, and use Grafana >= 8.3 and Prometheus >= 1.0.0 .

OpenMLDB provides Prometheus and Grafana configuration files for reference, see OpenMLDB mixin:

  • prometheus_example.yml: Prometheus configuration example, remember to modify the target address in node, openmldb_components and openmldb_exporter jobs for your environment

  • openmldb_dashboard.json: Grafana dashboard configuration for OpenMLDB metrics, import it two steps:

    1. Under the Grafana data source page, add the started Prometheus server address as the data source

    2. Under the dashboard browsing page, click the Import button, paste the dashboard ID 17843, or upload this json file directly

Understand Existing Monitoring Metrics#

Taking the OpenMLDB cluster system as an example, there are two type of metrics categorized by Prometheus pull jobs:

1. DB-Level metrics#

Exposed through the OpenMLDB exporter, the job_name=openmldb_exporter entry in the prometheus_example.yml:

  - job_name: openmldb_exporter
    # pull OpenMLDB DB-Level specific metric
    # change the 'targets' value to your deployed OpenMLDB exporter endpoint
    static_configs:
      - targets:
        - 172.17.0.15:8000

In detail three type of metrics:

  • Component status: table/nameserver/… information and status

  • Table status: database table related information, such as rows_count, memory_bytes

  • Deploy query response time

All metrics should visible already after a successful setup, excepting deploy query response time. Deploy query response time requires a extra global variable deploy_stats on, from OpenMLDB CLI:

SET GLOBAL deploy_stats = 'on';

The full DB-Level metrics can be listed through the following command:

curl http://172.17.0.15:8000/metrics

2. Component-Level metrics#

The related components of OpenMLDB (nameserver, tablet, etc), themselves as BRPC server, and expose [prometheus related metrics](https://github.com/apache/incubator-brpc/blob/master/docs/en/bvar .md#export-to-prometheus), you only need to configure the Prometheus server to pull metrics from the corresponding address. It corresponds to the job_name=openmldb_components item in prometheus_example.yml:

  - job_name: openmldb_components
    # job to pull component metrics from OpenMLDB like tablet/nameserver
    # tweak the 'targets' list in 'static_configs' on your need
    # every nameserver/tablet component endpoint should be added into targets
    metrics_path: /brpc_metrics
    static_configs:
      - targets:
        - 172.17.0.15:9622

The metrics of exposure are mainly

  • BRPC server process related information

  • Corresponding to the RPC method related metrics defined by the BRPC server, such as the RPC request count, error_count, qps and response_time

Metrics and help information can be shown through the following command (Note that the metrics exposed by different components will vary):

curl http://${COMPONENT_IP}:${COMPONENT_PORT}/brpc_metrics