How to analyze and visualize Singer metrics?

[edgarrmondragon]

(Related to work being done in #978)

Metric messages

Quoting from the Singer spec:

A Tap should periodically emit structured log messages containing metrics about read operations. Consumers of the tap logs can parse these metrics out of the logs for monitoring or analysis.

INFO METRIC: <metrics-json>

<metrics-json> should be a JSON object with the following keys:

This spec is missing some things. At least a timestamp that makes these records useful and shippable to systems like Prometheus or InfluxDB.

For example:

2022-09-28 01:20:06,717 INFO METRIC: {"metric_type": "timer", "metric": "http_request_duration", "value": 0.382852, "tags": {"stream": "continents", "endpoint": "", "http_status_code": 200, "status": "succeeded"}}

Other formats

One that's interesting is InfluxDB's line protocol:

myMeasurement,tag1=value1,tag2=value2 fieldKey="fieldValue" 1556813561098000000

It's easy enough for the user to convert metric logs to this format if they want to, so this would make for a nice guide on "How to aggregate and analyze Singer metrics?" or similar.