Drop filters

Drop Filters let you automatically discard noisy or irrelevant log lines, metrics, or spans before they count toward your ingest quota. Logs, metrics, or spans that match an active filter are never stored in Observe, so they:

  • cannot be searched,

  • do not trigger alerts,

  • never appear in dashboards, reports, or any other views.

Your monthly bill is calculated from Remaining observations, not the discarded Source observations.

Warning

You may filter out up to 50% of your average daily ingest volume (measured without Drop Filters enabled).

Requirements

To create a drop filter, you must either:

  • be an Admin (or a member of the Admin group), or

  • hold at least Edit access on the dataset where you’re adding the filter.

Without one of these privileges, the drop-filter option is unavailable.

Limitations

  • Drop filters can be attached only to source datasets or top‑level datasets (for example, Kubernetes Explorer/OpenTelemetry Logs, Tracing/Span Raw, or Metrics/OpenTelemetry).

  • Drop Filters DO NOT apply to data ingested via filedrop or poller (e.g., Add Data for AWS,filedrop‑based AWS integrations, or pollers for AWS and GCP).

  • Each source or top-level dataset can have up to 100 drop filters.

Create a drop filter

  1. Go to the Workspace settings page.

  2. Select Drop filters.

  3. Click New drop filter.

  4. Choose a targe dataset.

  5. Define one or more filter expressions that identify the noisy or irrelevant log lines you want to discard.

  6. Specify the drop rate (e.g., 100 %, 95 %).

  7. Click Save drop filter.

  8. Allow 1–2 minutes for the filter to become active.

To create or manage drop filters programmatically via Terraform, use the observe_drop_filter resource in the Observe Terraform Provider documentation.

Delete a drop filter

  1. Disable the filter.

  2. Click Delete to remove it permanently.