Health bar

A thin strip pinned above the breadcrumb on every Logs page shows three real-time metrics for the active project. Refreshes every 30 seconds.

What each chip means

INGEST. Number of events your server sent in the last 60 seconds. Computed from the per-minute rollup, so it reflects what actually landed in the database.

• Orange when there is activity
• Muted grey when zero (the server is quiet)

ERRORS. Number of events with severity error or fatal in the last 60 seconds.

• Muted grey when zero
• Amber when there are some
• Red when 10 or more in the last minute (your server is probably having a bad time)

LAST EVENT. How long ago the most recent event arrived for this project.

• Orange when fresh (under 2 minutes)
• Red when stale (over 2 minutes) - the server may have stopped sending
• "no events yet" if the project has never received an event

The right side of the bar shows a small refreshes every 30s note so you know the cadence.

Why it is useful

Ambient awareness without searching. Mods who keep the Logs page open in a tab can see at a glance:

• Whether the server is sending events at all
• Whether the error rate is spiking
• Whether ingest just stopped (the LAST EVENT chip turns red within ~2 minutes)

For incident response, the bar is often the first signal that something is wrong. A red ERRORS chip is "investigate now"; a red LAST EVENT is "the server crashed, ssh in".

What it is NOT

The health bar is not an SLA or status page. It is a local-to-this-project gauge. For platform-wide status (is the ingest pipeline healthy across all customers?) see the public status page.

The bar does not yet show:

• p50 / p99 query latency
• Per-dataset breakdowns
• Comparison to a baseline (last hour vs same hour yesterday)

Those are tracked as follow-up work.

When it shows nothing

If the project is brand new and has never received events, the bar shows zeros and "no events yet". This is the expected state for a freshly-created project. Send your first event (see Sending events) and within a minute the chips should populate.

If you see "no events yet" but you believe events ARE being sent, check:

• Are you looking at the right project? The chip in the top right shows the SCOPE.
• Is the ingest API token valid? An invalid token returns 401 without storing anything.
• Did the events return 400 from the API? Missing dataset or message causes that. See Sending events.

Per-project scope

The bar always shows metrics for the project shown in the breadcrumb / scope picker. Switching projects via the scope picker reloads the bar with the new project's numbers. Cross-project / cross-org leakage is impossible at the database layer.