Changelog

March 2026

Monitors rebuilt around the chart, charts inside alerts, every OpenTelemetry metric type, and far fewer noisy issues.

  • Monitors & alerts
  • Issues
  • Dashboards & metrics
  • Logs & traces
  • Integrations
  • Onboarding & billing
  • Self-hosting

Highlights

  • Monitors evaluate on schedule, re-notify on an interval you set, stop after a set count, and send alert and recovery emails.
  • Error and monitor alerts in Slack, Discord, email and PagerDuty include a trend chart and occurrence counts.
  • All OpenTelemetry metric types are stored: sums, histograms, exponential histograms and summaries.
  • Issues open only for error and warning patterns, and identical errors merge before an issue is created.
The monoscope monitors list, alerting monitors first.

Monitors & Alerts

  • Monitors that fire. Query monitors evaluate on schedule and notify. Set how often a monitor re-notifies and when it stops, and get alert and recovery emails.
  • Chart-first detail page. Threshold lines, a time picker, a details sidebar, and mute, resolve and deactivate actions. Alerts are now called Monitors across the app.
  • Charts in alerts. Slack, Discord, email and PagerDuty alerts include a trend chart and occurrence counts.
  • The monitors list puts alerting monitors first and shows short relative run times. Duplicate notification emails are gone.

Issues

  • Less noise. Issues open only for error and warning log patterns. Infrastructure noise is skipped, identical errors merge first, and regression alerts repeat less.
  • Issues page. Severity colors and icons, a short #N number for each issue that you can search by, an activity log, and syntax-highlighted code.
  • Log pattern issues show a 7-day volume chart, with event counts over 24 hours or 7 days.
  • Cleaner endpoints. Paths with IDs merge into templates such as /users/{id}, so the API catalog stops filling with duplicates.
  • Plain OpenTelemetry. The API catalog picks up standard OpenTelemetry HTTP spans, and shows sparklines with a Last Seen column.

Dashboards & Metrics

  • Every metric type. Sums, histograms, exponential histograms and summaries are stored, not only gauges, so CPU and memory from infrastructure agents show up. Metric queries are faster.
  • Endpoint dashboard. Overview, Errors and Latency tabs, with downstream calls at any depth.
  • Built-in dashboards have fewer tabs, table rows open the matching view, and timestamps and tags are easier to read.
  • Dashboards work on phones, and reordering widgets can no longer wipe another tab.

Logs & Traces

  • Waterfall. Hover tooltips, a time cursor and labeled stats. Clock skew is corrected, so child spans no longer start before their parents.
  • Merged patterns. Similar log patterns merge, with a badge showing how many.
  • The explorer, onboarding and settings work on phones.
  • A command palette, faster facets, and queries on missing columns return empty results instead of an error.

Integrations

  • Send telemetry with an x-api-key header as well as Authorization. The bundled collector accepts it from browsers.
  • Integrations live under Settings, Integrations.

Onboarding & Billing

  • Weekly reports. Redesigned emails, and the reports page shows each email as sent plus a live week-to-date preview.
  • Free projects see a banner and get emails as they approach and reach the daily limit.
  • Stripe checkout, and usage-based overage on the bring-your-own-storage plan.
  • Removed members can no longer open the project, and a project audit log records critical changes.

Self-Hosting

  • Docker Compose runs TimescaleDB on Postgres 18.
  • Telemetry is compressed after an hour and deleted after 30 days by default.
  • PAUSE_NOTIFICATIONS turns off monitor, error and endpoint notifications for the whole install.

Know what happened. Prove it.