Monitoring
Monitoring is a crucial part of database administration. By monitoring metrics, you can assess the state of the database, maintain the deployment without crisis, and diagnose problems when they occur.
For detailed metrics of GreptimeDB, please refer to the Metrics Detail section.
Start GreptimeDB
Please refer to the documentation to learn how to start GreptimeDB.
Export metrics to Prometheus
GreptimeDB supports exporting metrics to Prometheus. Before configuring export of metrics, you need to setup Prometheus by following their official documentation.
To scrape metrics from GreptimeDB, write a Prometheus configuration file and save it as prometheus.yml
:
global:
scrape_interval: 15s
scrape_configs:
- job_name: 'greptimedb'
static_configs:
# Assuming that GreptimeDB is running locally.
# The default HTTP port of 4000.
- targets: ['localhost:4000']
Start Prometheus using the configuration file. For example, bind-mount the configuration file when starting Prometheus using Docker:
docker run \
-p 9090:9090 \
-v $(pwd)/prometheus.yml:/etc/prometheus/prometheus.yml \
prom/prometheus
To avoid accidently exit the Docker container, you may want to run it in the "detached" mode: add the -d
flag to
the docker run
command.
Save metrics to GreptimeDB itself
You can also save metrics to GreptimeDB itself for convenient querying and analysis using SQL statements. This section provides some configuration examples. For more details about configuration, please refer to the Monitor metrics options.
Standalone
In standalone mode, you can simply use self_import
to export metrics.
The configuration looks like this:
[export_metrics]
enable=true
# The interval of writing metrics.
write_interval = "30s"
[export_metrics.self_import]
db = "information_schema"
The db
option specifies the database where metrics are saved. You can change it to a different database.
Distributed cluster
Configuration files need to be written for each component in the cluster.
Frontend
you can simply use self_import
to export metrics.
[export_metrics]
enable=true
# The interval of writing metrics.
write_interval = "30s"
[export_metrics.self_import]
db = "information_schema"
The db
option specifies the database where metrics are saved. You can change it to a different database.
Datanode and Metasrv
To export metrics for Datanode and Metasrv, you can use the remote_write
configuration:
[export_metrics]
enable=true
write_interval = "30s"
[export_metrics.remote_write]
url = "http://127.0.0.1:4000/v1/prometheus/write?db=system"
GreptimeDB is compatible with the Prometheus Remote-Write protocol. For more information, please refer to the Prometheus Remote-Write documentation.
Metrics Detail
You can check the output of curl http://<host>:<port>/metrics
by getting the latest metrics of GreptimeDB. We will add more documents of the metrics sooner.
Frontend
Key | Type |
---|---|
greptime_table_operator_ingest_rows | counter |
greptime_servers_error | counter |
greptime_servers_http_requests_total | counter |
greptime_servers_postgres_connection_count | gauge |
greptime_servers_mysql_connection_count | gauge |
greptime_query_merge_scan_regions | summary |
greptime_servers_http_sql_elapsed | summary |
greptime_query_optimize_physicalplan_elapsed | summary |
greptime_frontend_handle_sql_elapsed | summary |
greptime_http_track_metrics | summary |
greptime_query_create_physicalplan_elapsed | summary |
greptime_servers_mysql_query_elapsed | summary |
greptime_servers_http_requests_elapsed | summary |
greptime_query_execute_plan_elapsed | summary |
greptime_catalog_kv_get_remote | summary |
greptime_grpc_region_request | summary |
greptime_query_merge_scan_poll_elapsed | summary |
greptime_catalog_kv_get | summary |
greptime_table_operator_create_table | summary |
Datanode
Key | Type |
---|---|
greptime_opendal_bytes_total | counter |
greptime_servers_http_requests_total | counter |
greptime_opendal_requests_total | counter |
greptime_catalog_catalog_count | gauge |
greptime_catalog_schema_count | gauge |
greptime_opendal_requests_duration_seconds | summary |
greptime_http_track_metrics | summary |
greptime_servers_http_requests_elapsed | summary |
Meta
Key | Type |
---|---|
greptime_meta_create_schema | counter |
greptime_servers_http_requests_total | counter |
greptime_meta_create_catalog | counter |
greptime_meta_heartbeat_connection_num | gauge |
greptime_meta_txn_request | summary |
greptime_meta_kv_request | summary |
greptime_meta_create_schema | summary |
greptime_meta_create_catalog | summary |
greptime_meta_handler_execute | summary |
greptime_servers_http_requests_elapsed | summary |
greptime_http_track_metrics | summary |
greptime_meta_procedure_create_table | summary |
greptime_grpc_region_request | summary |