Prometheus¶
Metrics¶
Metric |
Description |
Service |
|---|---|---|
dataclay_inmemory_objects |
Number of objects in memory |
backend, client |
dataclay_loaded_objects |
Number of loaded objects |
backend |
dataclay_stored_objects |
Number of stored objects |
backend |
dataclay_inmemory_misses_total |
Number of inmemory misses |
backend, client |
dataclay_inmemory_hits_total |
Number of inmemory hits |
backend, client |
Deploy dataClay with Prometheus¶
Run dataClay with Prometheus:
Note
This example is available in GitHub.
docker compose up -d
The metadata-service and backends will post their metrics to the 8000 port.
Prometheus is configured to scrape this port to pull the metrics.
Access Prometheus at http://localhost:9090. You can query the metrics defined above.
Deploy dataClay with Prometheus Pushgateway¶
Run dataClay with Prometheus Pushgateway:
Note
This example is available in GitHub.
docker compose up -d
The metadata-service and backends will push their metrics to the pushgateway at the 9091 port.
The client.py can also push metrics using the pushgateway:
export DATACLAY_METRICS=true
export DATACLAY_METRICS_EXPORTER=pushgateway
export DATACLAY_METRICS_HOST=localhost # the default
export DATACLAY_METRICS_PORT=9091
python3 client.py
Access the Pushgateway at http://localhost:9091 and Prometheus at http://localhost:9090.
Note
When using pushgateway, a new Python thread will run to push the metrics every 10 seconds (default).