# Monitoring



Krenalis monitors its own internal service metrics, and makes them available under the `/metrics` path.

You can control whether this endpoint is available by using the `KRENALIS_PROMETHEUS_METRICS_ENABLED` environment variable. By default, this variable is set to `false`, so the `/metrics` endpoint is disabled unless you explicitly enable it. In Docker Compose installations, `KRENALIS_PROMETHEUS_METRICS_ENABLED` is set to `true` by default. This means the `/metrics` endpoint is automatically enabled in that setup.

## Metrics

| Name                                         | Type         | Labels                       | Description                                                              |
|----------------------------------------------|--------------|------------------------------|--------------------------------------------------------------------------|
| `krenalis_db_acquired_conns`                 | Gauge        | –                            | Current number of connections in use by the database connection pool.    |
| `krenalis_db_max_conns`                      | Gauge        | –                            | Configured maximum number of simultaneous database connections.          |
| `krenalis_db_acquire_duration_seconds_total` | Counter      | –                            | Cumulative seconds spent acquiring connections from the database pool.   |
| `krenalis_db_acquire_count_total`            | Counter      | –                            | Total number of successful database connection acquisitions.             |
| `krenalis_db_new_conns_count_total`          | Counter      | –                            | Total number of newly created database connections (connection churn).   |
| `krenalis_lambda_errors_total`               | CounterVec   | `type`                       | Total number of Lambda errors, categorized by error type.                |
| `krenalis_lambda_duration_seconds`           | Histogram    | –                            | Duration of successful Lambda executions, in seconds.                    |
| `krenalis_lambda_records_total`              | Counter      | –                            | Total number of input records processed by successful Lambda executions. |
| `krenalis_sender_queue_available`            | GaugeVec     | –                            | Number of available events in the event queue.                           |
| `krenalis_sender_queue_wait`                 | HistogramVec | `connector`,<br>`connection` | Time spent waiting in the event queue (in seconds).                      |

### Notes

- The total number of successful Lambda executions is provided by the metric `krenalis_lambda_duration_seconds_count`, which is part of the` krenalis_lambda_duration_seconds` histogram. It represents the total number of observations (i.e., completed executions) recorded in the histogram.
- Possible values for the `type` label in `krenalis_lambda_errors_total`:  
  `network`, `lambda_internal`, `function_not_found`, `serialization`, and `function_exec`.
- Buckets defined for `krenalis_lambda_duration_seconds`: `0.1`, `0.5`, `1`, `2.5`,  and `5` (in seconds)
- Buckets defined for `krenalis_sender_queue_wait`: `0.005`, `0.01`, `0.025`, `0.05`, `0.075`, `0.1`, `0.15`, `0.2`, `0.3`, `0.5`, `0.75`, `1.0`, `2.0` (in seconds)

## Prometheus

To monitor these metrics with [Prometheus](https://prometheus.io/), configure it with the following scrape job:

```yaml
scrape_configs:
  - job_name: 'krenalis'
    static_configs:
      - targets: [ "127.0.0.1:2022" ]
```

## Grafana

After setting up Prometheus to collect the metrics, configure [Grafana](https://grafana.com/oss/grafana/) to display them in dashboards.