Service Level Agreement

1. Scope and Applicability

This Service Level Agreement ("SLA") describes the uptime commitments, credit calculation methodology, and support response times for paid Partikl plans.

This SLA is incorporated into and governed by the Terms of Service. Defined terms used here have the meanings given in the Terms of Service.

This SLA applies to: Pro, Team, and Enterprise plans.

This SLA does not apply to:

• Free tier accounts (best-effort service, no commitments)
• Features explicitly marked as "Beta" in the dashboard or documentation
• Third-party addons and custom workers
• BYOS (Bring Your Own Storage) availability
• Scheduled maintenance windows (with advance notice)
• Force majeure events as defined in Section 6

2. Uptime Commitments

2.1 Monthly Uptime Targets

Monthly uptime percentage is calculated as:

Uptime % = ((Total minutes in month − Downtime minutes) / Total minutes in month) × 100

A standard month is calculated as 43,200 minutes (30 days).

2.2 What Counts as Downtime

Downtime is defined as a period during which the following services are completely unavailable for your account, measured in continuous minutes:

Partial outages: If an outage affects a subset of regions or namespaces, downtime is calculated proportionally based on the percentage of your namespaces affected during the outage window.

2.3 What Does Not Count as Downtime

The following are explicitly excluded from downtime calculations:

• Scheduled maintenance with advance notice per Section 4
• BYOS storage unavailability — your storage provider's outages
• Third-party addon failures — addons operated by third parties
• Custom worker or webhook endpoint failures — your infrastructure
• Force majeure events as defined in Section 6
• Account suspension due to policy violations or non-payment
• Free tier service interruptions (no SLA applies)
• Beta feature outages — features explicitly marked Beta
• Client-side issues — network problems on your end, DNS misconfiguration
• DDoS attacks that exceed industry-standard mitigation capacity
• Outages caused by your pipeline configuration — infinite loops, resource exhaustion due to misconfigured workflows

3. Credit Calculation and Issuance

3.1 Credit Tiers

Credits are calculated based on the actual downtime minutes recorded in a given calendar month and the monthly fee paid for your plan (base subscription, excluding overage charges).

For Team plan (99.95% target):

3.2 Credit Cap

Credits issued under this SLA are capped at 50% of the monthly base subscription fee for the affected billing period, regardless of the severity or duration of the outage.

Credits are applied to your account balance and deducted from future invoices. Credits are not redeemable for cash and expire 12 months after issuance if unused.

3.3 Credit Issuance Process

Automated issuance: For outages tracked and confirmed on our status page (status.partikl.io), credits are calculated automatically and applied to affected accounts within 5 business days of the end of the billing period. No claim is required.

Manual claim: For outages not captured by our automated monitoring (for example, isolated namespace issues), you may submit a claim:

1. Navigate to Dashboard → Billing → SLA Credit Request
2. Provide: incident date/time, affected namespace(s), and a description
3. Submit within 30 days of the incident
4. We will respond within 5 business days

We verify claims against our internal ClickHouse metrics and infrastructure logs. Decisions are final but may be appealed once by contacting sla@partikl.io.

3.4 Credits as Sole Remedy

4. Scheduled Maintenance

We perform scheduled maintenance to improve reliability, apply security patches, and upgrade infrastructure. We minimize the frequency and duration of maintenance windows.

Scheduled maintenance is announced via:

• Email to account owner and billing contacts
• Dashboard banner
• Status page (status.partikl.io)

Emergency maintenance required to address active security vulnerabilities or cascading failures may be performed without advance notice. We will communicate the reason and post a post-mortem within 48 hours.

Scheduled maintenance windows do not count as downtime.

5. Support Response Times

Response times are measured from receipt of your support request during our support hours. Response does not guarantee resolution — it indicates we have acknowledged your request and begun investigation.

Priority classification:

Support response time SLA is separate from uptime SLA. Failure to meet support response times does not generate uptime credits.

6. Force Majeure

Partikl is not liable for downtime caused by circumstances outside our reasonable control, including but not limited to:

• Natural disasters, earthquakes, floods, fires
• War, armed conflict, terrorism, civil unrest
• Government actions, regulatory orders, sanctions
• Pandemic or public health emergencies
• Internet backbone failures or BGP routing disruptions
• Upstream infrastructure provider failures (datacenter power, physical damage)
• Large-scale coordinated cyberattacks exceeding industry-standard defenses
• Failures in third-party services we depend on (payment processors, DNS registrars)

We will communicate force majeure events promptly on our status page and via email. We commit to resuming normal operations as quickly as reasonably possible and will provide a post-mortem for events lasting more than 2 hours.

7. Status Page and Incident Communication

Our public status page at status.partikl.io provides:

• Real-time service status for all components
• Active incident updates (updated at least every 30 minutes during incidents)
• Historical uptime data (rolling 90 days)
• Post-mortem reports for significant incidents

Incident communication during active outages:

We use our status page as the authoritative source of truth for uptime calculations. If our monitoring shows a longer outage than your own records, we will apply the longer duration for credit calculations.

8. Measurement and Verification

Uptime is measured from multiple monitoring locations using synthetic probes that test:

• API endpoint availability and response validity
• Dashboard load and authentication flow
• Asset upload acceptance
• Asset delivery from primary edge locations

Monitoring is performed at 1-minute intervals. An outage begins when 2 consecutive probes from at least 2 monitoring locations fail. An outage ends when 3 consecutive probes succeed from all locations.

Internal metrics from our ClickHouse infrastructure observability cluster serve as the authoritative source for SLA calculations. These metrics are available in aggregated form upon request by Enterprise customers.

9. Enterprise SLA

Enterprise plans receive a customized SLA negotiated as part of the Enterprise Agreement. Custom SLAs may include:

• Higher uptime targets (up to 99.99%)
• Custom maintenance windows agreed in advance
• Dedicated support contacts and escalation paths
• Financial remedies beyond standard credit tiers (by negotiation)
• Access to infrastructure metrics and audit logs
• Custom monitoring integrations

Contact enterprise@partikl.io to discuss Enterprise SLA options.

10. Changes to This SLA

Changes to uptime targets or credit calculations are material changes and will be communicated with 30 days notice before taking effect.

Changes to support response times, maintenance notice periods, or measurement methodology are standard changes requiring 14 days notice.

View the full change history in our Legal Changelog.

This document was last updated on April 6, 2026 (v1.0).
The Legal Changelog records all historical changes.