Reliability

Service status

The live operational picture: every system across every region, the last few incidents, and how to subscribe to updates.

Last updated · May 14, 2026

Overview

This page is the source of truth for margininfo's operational status. Live indicators below are refreshed automatically every minute. Incident history and post-mortems are appended whenever an incident is closed.

If something feels wrong but this page is green, please open the in-product help and tag #trust-status, our on-call team will investigate within minutes.

Maintenance windows

We schedule infrastructure maintenance during low-traffic windows: 02:00–04:00 UTC on the second Tuesday of each month. Planned maintenance appears in the incident history below at least 72 hours in advance.

Customer-facing degradations during scheduled windows are minimised, most maintenance is invisible to users.

Service status

Live operational status, by service and region. All systems below are reporting healthy. Uptime is rolling 90 days.

All systems operational checked · 2 min ago

Marketing site global

99.999% Operational
App (web) US + EU

99.982% Operational
API US + EU

99.978% Operational
Agent runtime US + EU

99.961% Operational
Investigations (jobs) US + EU

99.953% Operational
Connectors, Shopify global

99.991% Operational
Connectors, Meta + Google Ads global

99.973% Operational
Connectors, Amazon global

99.962% Operational
Billing + Stripe global

100.000% Operational
Dashboards + reports US + EU

99.984% Operational

Incident history

A timeline of the last few notable incidents. Detailed post-mortems are linked from the monthly reliability digest sent to workspace admins.

May 2, 2026 Minor · 37 min

Elevated latency on Meta Ads connector backfills

Backfill jobs older than 14 days queued behind upstream rate limits. New runs and live syncs were unaffected. Resolved by widening the backoff window and adding parallel partitions.
April 18, 2026 Minor · 11 min

Brief degradation of investigation queue (EU)

EU agent runtime saw a 2× queue depth after a node group rolled. Auto-scaler kicked in; no investigations were lost or duplicated.
March 26, 2026 Scheduled · 8 min

Scheduled maintenance, database failover drill

Planned read-replica promotion drill. App was read-only for ~8 minutes during the East-US window (02:00–02:08 UTC). No data impact.

Overview

Maintenance windows

Service status

Incident history

Elevated latency on Meta Ads connector backfills

Brief degradation of investigation queue (EU)

Scheduled maintenance, database failover drill

Subscribe to status