Service level agreement

Taska targets 99.5% monthly uptime for the core authenticated surfaces (Today, Tasks, Calendar, Stats, Bookings) and the public booking pages (/book/:slug).

99.5% means we may be unavailable up to ~3 hours 39 minutes per calendar month before the SLA is breached.

• HTTP 500-class errors on the API for > 5 consecutive minutes, as measured by our external uptime monitor.
• Sign-in failure rate > 5% over a 15-minute window.
• Task extraction worker stopped processing (no jobs advancing) for > 30 minutes during business hours (08:00–22:00 UTC).

• Scheduled maintenance windows announced 24+ hours ahead on /status and via email.
• Outages caused by upstream providers we depend on (AWS, Stripe, Cognito, Gmail/Outlook OAuth providers, Bedrock). We pass through their reported incidents on our status page but can't credit for them.
• Force majeure events (regional internet outages, large-scale cloud-provider failures, etc.).
• Customer-caused issues (revoked OAuth, exhausted Bedrock personal cap, exceeded rate limits).

When we miss the 99.5% target in a calendar month, paying customers can request a service credit:

Credits are applied to your next invoice and don't roll over past 90 days. Free-tier accounts don't qualify (no fee to credit).

Email hello@enrevia-taska.com within 30 days of the affected month. Include your account email and the dates of the impact. We'll cross-check against our monitoring + status-page postmortems and confirm within 5 business days.

Live operational state is at /status. Material incidents (customer-data impact) trigger an email to affected accounts within 72 hours per the security policy, with a postmortem published within 5 business days.