Real-time uptime and incident history for Astrolium. The live dashboard auto-refreshes every 30 seconds at status.astrolium.app; this page is the human-readable explanation. Anything Astrolium misses the SLA on, we credit your account automatically. SLA target: 99.9% monthly. April 2026 actual: 99.96%. Last incident: 1 day ago. See pricing for tier-specific SLA terms, security for the underlying infrastructure, and terms for the contractual framing.
Components — service by service
All 8 monitored components, with current state and 30 day uptime:
- Web app — Operational, 99.99% uptime. All regions nominal.
- Ephemeris API — Operational, 99.97% uptime. p99 latency 41 ms. See the engineering walkthrough for what's inside.
- Chart rendering — Operational, 100% uptime. 0 errors in the last 24 hours.
- PDF export — Operational, 99.94% uptime. Slight retry uptick on April 18, resolved.
- Email delivery — Operational, 99.98% uptime. Postmark, EU region.
- Authentication — Operational, 100% uptime. All providers nominal.
- Sync (clients) — Degraded, 99.71% uptime. Elevated latency in EU region; under active monitoring.
- Marketing site — Operational, 100% uptime. CDN cached.
Last 90 days at a glance
Of the last 90 days, 86 were fully operational, 3 had minor degradation, and 1 had a major incident (the March 9 authentication outage, detailed below). The 90 day uptime grid is rendered on the live status page; this page records the underlying events in plain language.
Incident history
Plain language postmortems for everything over 3 minutes. Anything longer than 30 minutes gets a full root cause analysis.
April 18, 2026, 14:32 to 14:51 CEST — elevated PDF export retries
Severity: Minor. Duration: 19 min. Status: Resolved.
A memory leak in the headless PDF renderer caused PDF export jobs to retry once before succeeding. No exports failed; users may have noticed exports taking 10 to 15 seconds longer than usual. Patched and deployed within 19 minutes. Underlying cause: a font subsetting cache wasn't being cleared between jobs at high concurrency.
March 28, 2026, 09:14 to 10:46 CEST — client sync slow in EU region
Severity: Minor. Duration: 1 h 32 min. Status: Resolved.
Client roster sync was slow (5 to 15 seconds) for users in the EU region while the primary database underwent an unscheduled storage migration triggered by a noisy neighbor. Read replicas remained nominal; chart rendering was unaffected. Migration completed and primary returned to normal latency.
March 9, 2026, 21:08 to 22:55 CEST — authentication brief outage
Severity: Major. Duration: 1 h 47 min. Status: Resolved.
Login was unavailable for roughly 110 minutes during a misconfigured rolling deploy. Users already signed in were unaffected. New logins and password resets failed during the window. Root cause: the new image was missing an environment variable the auth service expects on cold start. Mitigated by a manual rollback; the deploy pipeline now blocks rollouts that fail a synthetic login probe.
February 14, 2026, 06:11 to 06:14 UTC — Cloudflare regional blip
Severity: Minor. Duration: 3 min. Status: Resolved upstream.
Cloudflare incident in the AMS region. Marginal effect on EU users (some retried connections); fully transparent to most. No action on our side.
Get notified when something breaks
Subscribe to status updates by email at status@astrolium.app (subject: Subscribe) or via RSS. Astrolium posts within 5 minutes of detection and writes postmortems for anything longer than 30 minutes.
For the SLA terms by tier see pricing. For the security posture behind the infrastructure see the security page. For the engineering details of the ephemeris service see the writeup.