What's new

SAML signature-bypass vulnerability via goxmldsig patched; npm advisory fixes applied across frontend dependencies.

Nullable alert labels and JSONB fields now guarded — prevents crashes when alert sources send partial or missing data.

Brand logo inlined so it renders correctly in PostHog session replays.

Go upgraded to 1.25.11 and golangci-lint updated to v2 for cleaner, faster builds.

Timezone badge on Add Layer, Edit Layer, and Create Override modals — shows the schedule timezone and current time in that zone.

"Use schedule midnight" quick-fill — one click fills the next midnight in the schedule's timezone, expressed as your local time.

Live timezone hint below Rotation Start — translates the time you typed into the schedule's timezone in real time.

Shift end in who's on call card — shows countdown and schedule-tz wall clock when it differs from your browser timezone.

Gantt calendar timezone footer — shows the schedule timezone and current time below the calendar grid.

App version shown in sidebar footer (e.g. v0.18.0) so you always know what's running without SSHing into the server.

Schedule datetime inputs — all datetime-local fields now correctly use your browser's local timezone instead of UTC; rotation starts and overrides no longer silently shift time on save.

Mark unavailable — fixed "Invalid request parameters" error when submitting unavailability dates.

Background tab errors — polling now pauses when the tab is hidden and resumes on focus, preventing "Failed to fetch" errors after the tab has been idle.

Incident attachments - upload, view, download, and delete files directly from the incident detail page.

Post-mortem template selector - choose a template when generating post-mortems from the UI.

Routing rule criteria - regex and annotation matching now supported in routing rule conditions.

PagerDuty import - EU region support added to the import wizard.

Slack HTTP Events API - migrated from Socket Mode to HTTP Events API for more reliable event delivery.

AI cost tracking - OSS foundation for per-operation cost tracking (BYO API key).

Routing rule suppress toggle, priority default, and optional webhook fingerprint.

Holiday dates shifting due to timezone handling.

Slack incident card now refreshes and posts confirmation after reaction-based ack/resolve.

`cost_usd` returned as a flat field on PostMortem API response (was incorrectly wrapped).

SQL migrations embedded as `fs.FS` at compile time - removes the `./migrations` file path dependency, enabling clean Docker and binary deployments.

`enterprise.Hooks.CustomFieldsHandler` - Pro binary registers custom field routes; OSS returns 402.

`enterprise.Hooks.UIProvider` - Pro binary serves its own embedded frontend; OSS behaviour unchanged.

`commands.SetEnterpriseHooks()` - called by the Pro binary before `Execute()` to inject all hooks.

Helm chart publishing to `charts.fluidify.ai` on each release.

Helm install URL in README.

Public status page at `/status` - no login required. Shows active incidents with live severity indicators and incidents resolved in the last 7 days.

Status page URL shown in Settings → System for easy sharing with your team.

Incident detail redesign - severity-tinted header with a colour wash calibrated to urgency.

Left severity stripe on the incident title for instant visual context.

Sticky compact bar appears when scrolling past the title.

Properties panel with collapsible sections, richer field editing, and save-flash feedback.

Anthropic Claude support - connect with your own API key (Claude 3.5 Sonnet by default).

Ollama support - fully local LLM inference; point Regen at your Ollama endpoint and model.

AI provider switcher in Settings → System → AI Provider. All three power incident summaries, post-mortem drafts, and handoff digests.

Existing OpenAI users continue to work with no config changes.

In-app announcements - dismissible banner for admins when there are release notes or important updates, fetched automatically in the background.

Anonymous aggregate telemetry (incident counts, features enabled, instance size). No incident content or user data. Opt out with `REGEN_NO_TELEMETRY=1` or Settings → System.

Full incident lifecycle - triggered → acknowledged → resolved with immutable, append-only timeline.

Create incidents from alerts, Slack (`/incident new`), the UI, or REST API.

AI-generated post-mortems from incident timeline, Slack thread, and linked alerts (BYO OpenAI key).

Post-mortem templates - define required sections and guiding questions per team.

On-call schedules with multi-layer rotations, daily/weekly/custom shift types, timezone support, and overrides.

Escalation policies - multi-step with configurable timeouts. Escalates automatically on no-ack.

Shift notifications - incoming on-call engineer notified at handoff via Slack or Teams DM.

Slack integration - auto-creates incident channels, `/incident ack`, `/incident resolve`, interactive buttons.

Microsoft Teams - Adaptive Cards with full two-way sync, bot commands, incident channel creation.

Telegram - notification-only (severity-coded alerts, AI summary delivery).

Alert sources: Prometheus/Alertmanager, Grafana, AWS CloudWatch, generic webhooks.

Alert deduplication - fingerprint-based grouping prevents duplicate incidents.

One-click Grafana OnCall migration - imports schedules, escalation policies, and users in one step.

Opsgenie schedule and policy import.

Docker Compose quickstart - running in under 5 minutes.

Helm chart for Kubernetes with TLS ingress, Sentinel-mode Redis HA, and Patroni Postgres HA.

SAML / SSO - free forever, no enterprise gating (Okta, Azure AD, Google Workspace).

Non-root container (UID 1001), read-only filesystem, all Linux capabilities dropped.

Webhook ingestion p99 < 10 ms. Tested at 3,917 RPS peak with zero 5xx errors.

Fluidify MCP Server (beta) - external agents can read incidents, query timeline, and check on-call natively.