Skip to main content

TODO: ⚠️ This page is outdated 🚧

Get the insights you need about how your system is used.

Use an analytics storage​

All request scoped log records have a trace_id. Using the trace_id, join and aggregate the records together as you like. The easiest way to do so is to ingest the streams you need into an analytics storage. For example, you can do so by using the format "otel" and run an Otel collector that ingests the logs into your ClickHouse. Don't forget to make sure, your streams are configured to redact, remove or anonymize sensitive values and PII.

Examples​

Learn here about how billing relevant Zitadel Cloud metrics are producible to get some inspiration.

Active Users​

To calculate active users over a period of time, we need to distinct count the anonymized user ids in this period.

Example Stream Configuration​

Types:
- Type: "Authentications"
  Enabled: true
  Format: "otel"
  SensitiveFields: "remove"
  PersonallyIdentifiableInformation: "anonymize"
- Type: "Authorizations"
  Enabled: true
  Format: "otel"
  SensitiveFields: "remove"
  PersonallyIdentifiableInformation: "anonymize"

Example ClickHouse SQL​

SELECT toMonth(Timestamp) AS month,
       COUNT(DISTINCT path(Body, "Stream", "AuthZ", "UserID")) AS active_users
FROM otel_logs
WHERE path(Resource, "service.name") = "zitadel"
  AND (
    (
        path(Body, "Stream", "AuthN", "Success") = true
        AND path(Body, "Stream", "AuthN", "Method") IN ("OIDCAccessToken", "OIDCRefreshToken", "SessionAPI", "SAMLResponse")
    )
    OR (
        path(Body, "Stream", "AuthZ", "Success") = true
        AND path(Body, "Stream", "AuthZ", "UserID") != ""
    )
  )
GROUP BY month

TODO: Management Requests​