For the complete documentation index, see llms.txt. Markdown versions are available by appending .md to documentation URLs.

Changelog

Follow us on LinkedIn
v0.150.0November 16, 2026

⚠️ Action required (API users): v1 alert rules APIs removed

Alert rules now use a single schema, v2alpha1. From this release, the v1 rules APIs are removed. Your existing rules are converted to v2alpha1 automatically when SigNoz starts after the upgrade.

Who needs to act:

  • Scripts using the old endpoints: anything calling /api/v1/rules, /api/v1/testRule, or GET /api/v2/rules must switch to the replacement endpoints.
  • Scripts using the v2 endpoints: create, replace, and test requests must include "schemaVersion": "v2alpha1".
  • Anything that reads rules: rules now come back in the v2alpha1 shape.
  • Terraform users: no action is needed on signoz_rule. Move off the older signoz_alert resource.

For UI-only users, no action is needed.

Read the v0.150 upgrade guide before upgrading.

Alert Rules list moved to the v3 API, with new search and saved views

The Alert Rules list now uses the new /api/v3/rules API. Search, filtering and pagination now run server-side, so the browser only loads the page you’re viewing.

The search bar takes filters such as name, severity, alert type, rule type, created by, and labels, with autocomplete as you type.

We’ve also added saved views, which sit alongside the list. You can build views around your apps, your infrastructure, or any logic you want, and they are shared with your organization. You also get four built-in views: My alert rules, All, Firing, and Disabled.

Traces Explorer gets a new span drawer and infinite scroll

Reading spans in the Traces Explorer list is now easier. Clicking a row opens the span in a side drawer. You no longer have to open the trace page and come back to check the next span.

The drawer shows the same overview, events, and logs tabs as the trace page. Use the up and down buttons or the arrow keys to step through the rows. Click outside the table to close it.

The list view now loads spans as you scroll, 100 at a time, until it reaches the end of the results or the query limit. The Previous and Next buttons are gone.

Hovering a row shows an icon to open the full trace in a new tab.

Self-hosted Jira added to the Jira alert channel

You can now send SigNoz alerts to self-hosted Jira. This covers Jira Server and Data Center, version 8.14 or newer.

When creating a Jira channel, choose where your Jira runs. Self-hosted uses a personal access token. Cloud still uses your email and API token. Existing Jira Cloud channels keep working without changes.

If you manage channels through the API, the Jira spec has changed. deployment is now required, and credentials sit under auth.

Saved views and column options in AI Observability

You can now save a view in the AI Observability Explorer. Come back to it any time to monitor the same data or continue an investigation, without rebuilding it.

The list view also gets a column picker. You can choose which columns to show and arrange them the way you want. Your columns are saved with the view and carried in the URL.

Pie charts group small slices into Other

Pie charts with many groups are now easier to read. Slices under 1% of the total are grouped into a single Other slice. A pie also shows at most 15 slices.

The legend still lists every series. Hover over Other to see what’s inside it. Click it to pin the list, then click a member to drill into that series.

This happens automatically on the new dashboards, with no panel setting to change.

⚠️ Action required (OSS users): Plan for extra capacity before upgrading

Log bodies are moving to native JSON storage. From this release, logs are dual-ingested. The collector writes each log body twice: to the existing body column and to the new JSON column. Your dashboards, alerts, and saved views keep working unchanged.

Dual ingestion needs more resources. Plan for additional collector CPU and memory, ClickHouse CPU, and log disk space before you upgrade.

The migration guide has a query to estimate the disk increase.

Bug Fixes

  • JSON log bodies and object fields now render properly in every logs view, instead of [object Object] or a crash.
  • HAVING on formulas is now applied; it was previously accepted and ignored.
  • Time-shifted queries no longer return wrong data points when part of the time range is served from the cache.
  • Logs and traces charts with a group-by and a limit now show the correct top groups. Previously, cached results could show the wrong groups or partial series.
  • The Alert Details page no longer crashes when the URL has an invalid relativeTime. It now falls back to the default 30-minute range.
  • Max lines per row in Logs now clamps to 1–10.

Maintenance

  • Dashboards have a copy-ID button next to the title.
  • trace_id, span_id, and other high-cardinality fields can no longer be added as quick filters. Their values are unique per span or log, so the filter had nothing to pick from. They are removed from existing quick filters. You can still filter by them in the search bar.