Skip to contentManaged Kafka is live: Apache Kafka on AWS for half the cost of MSK

Monitoring

Analytics portal

Monitor your Artie pipelines on the Analytics page. View rows synced, operation types, table distribution, ingestion lag, and more with filtering by pipeline and table.

Once a pipeline is running, its metrics show up on the Analytics page. To alert on a configured condition, create a custom monitor.

Available filters

For each of these metrics, you also have the ability to set a filter:

  1. Pipeline (All or a specific pipeline)
  2. Specific tables (This is available if you specify the pipeline)
  3. Whether to include or exclude backfills
  4. Mode (Replication or history)

Metrics

The Analytics page groups its charts into two sections:

  • Volume: rows synced over time, rows by table (or by pipeline), and rows by operation (create, update, delete, backfill).
  • Freshness: lag (switch between reader, ingestion, and total lag) and row lag.

Below those sections, every pipeline shows flush time and flush count. PostgreSQL pipelines also show replication slot size. Snowflake pipelines show warehouse execution and queued time, and BigQuery pipelines show bytes processed.

Analytics

Pipeline Postgres to RedshiftTables AllMode Replication + HistoryBackfills IncludedPast 7 days

Volume

How much data we're moving

Rows synced

214.6M

Past 7 days

Rows by table

214.6M

  • orders81.5M38%
  • events57.9M27%
  • customers36.5M17%
  • payments23.6M11%
  • invoices15.0M7%

Rows by operation

214.6M

  • Inserts90.1M42%
  • Updates109.4M51%
  • Deletes6.4M3%
  • Backfills8.6M4%

Freshness

How current your data is

Lag

1.05 minutes

Latest

Row lag

33.59K

Latest

Flush time

1.54 seconds

Latest

Flush count

9,802

Past 7 days

  • reason: time24
  • reason: rows34
  • reason: size1

Replication slot

55.25 MB

Latest

Sample data. Point at any chart, or focus one and use the arrow keys, to read the same hour on every chart.

Query metrics with the API

The same metrics are available from GET /analytics with an API key. Pass a time window, the datasets you want, and optionally a pipeline:

Bash
curl -sg -H "Authorization: Bearer $ARTIE_API_KEY" \
  "https://api.artie.com/analytics?pipelineUUID=5fc762a9-d1ea-4dde-a986-fc46c26a2dd3&from=2026-10-30T22:00:00Z&to=2026-10-30T23:00:00Z&datasets[]=readerLag&datasets[]=ingestionLag"
JSON
{
  "datasets": {
    "readerLag": {
      "unit": "ms",
      "err": null,
      "data": [{ "dimensions": {}, "points": [{ "time": 1793401200000, "value": 120 }] }]
    },
    "ingestionLag": {
      "unit": "ms",
      "err": null,
      "data": [{ "dimensions": {}, "points": [{ "time": 1793401200000, "value": 610.34 }] }]
    }
  }
}

Each point’s time is a Unix timestamp in milliseconds, and value is in the dataset’s unit. Check err on every dataset: one dataset can fail while the others succeed.

DatasetAnalytics chart
rowsSyncedByTime, rowsSyncedByOperation, rowsSyncedByDBRows synced, rows by operation, rows by table (or by pipeline)
ingestionLagLag (ingestion): time from Artie receiving a change to writing it to your destination
readerLagLag (reader): how far behind Artie is reading your source
rowLagRow lag
flushTime, flushCountDestination flush duration and count
replicationSlotSizePostgres replication slot size
snowflakeExecutionTimeAvg, snowflakeExecutionTimeMax, snowflakeQueuedOverloadTimeAvg, snowflakeQueuedOverloadTimeMaxSnowflake warehouse execution and queued time
bigQueryUsageBigQuery bytes processed

Dataset names are case-sensitive, and unknown names are ignored.

Total lag (source to destination) is readerLag plus ingestionLag at each timestamp. Ingestion lag is only reported when Artie flushes, so treat a missing point as 0, which is how the Analytics page draws its total lag chart.

API keys scoped to one environment must pass pipelineUUID. Requests without it return 400, and a pipeline from another environment returns 404.