To our merchants,

Supercycle experienced outages over the last day or so that disrupted your operations. We know how much you depend on us to deliver stability, and downtime directly impacts your sales, your customers, and your reputation. For this, I want to offer a sincere and unreserved apology.

What happened

The root cause came down to how our database connection pool was configured relative to the workloads we were running. Webhooks can take 10–20 seconds to process. Under load, those long-lived connections were competing with standard web requests for the same limited pool of database connections. As the pool became saturated, incoming requests had to wait in line, which caused request latency to climb.

This slowdown cascaded into our health check endpoints, which our cloud orchestration platform (Cloud66) uses to confirm pod health. Because those checks were occasionally delayed past their timeout threshold, Kubernetes interpreted the pods as unhealthy and restarted them. The result was a restart loop, not because the pods were fundamentally broken, but because the health checks were being starved of resources at the database layer.

In hindsight, this is exactly the type of subtle performance tuning and monitoring that comes with managing your own database. A fully managed solution would have automatically handled connection pooling, isolation of long-lived queries, and scaling under load.

What we’re doing next

This incident occurred just before major upgrades we already had in our pipeline. We are accelerating those efforts and implementing additional safeguards to prevent recurrence:

Status page: We’ve launched status.supercycle.com to provide real-time transparency on uptime, incidents, and planned maintenance.
Server upgrades (multi-region HA): In the coming days, we will migrate to a multi-region, high availability setup for our web servers. This will significantly reduce single points of failure and provide resilience against regional outages.
Database migration (PlanetScale): We are moving from a Supercycle-managed database to PlanetScale’s fully managed platform, the same team trusted by companies like Etsy and Intercom. This migration will be fully managed for zero-downtime and will dramatically improve scalability and resilience.
Telemetry & monitoring: We’re rolling out stronger observability and telemetry tooling, so we can detect anomalies faster and address them before they escalate.
On-call & alerting: A more robust alerting system and on-call policy are being formalised, ensuring rapid response and accountability in any future incident.
Business continuity: We’re implementing both a business continuity policy and a structured incident management framework to improve crisis response and communication.

Balancing speed and stability

Supercycle is scaling rapidly, and much of our energy has been on building and shipping. But this incident has made clear that we need to dedicate more headspace and discipline to stability. Growth demands not just speed, but reliability. Going forward, we are intentionally rebalancing so that stability is treated as a first-class priority alongside innovation.

Our commitment

The changes underway are more than short-term fixes. Multi-region high availability, PlanetScale’s managed database, and our strengthened monitoring and incident processes form a new foundation for long-term reliability. These steps will ensure we can support your businesses with the stability you deserve as we grow together.

Thank you for your patience and for trusting us to be a critical part of your business. I don’t take that lightly, and I’m committed to making sure Supercycle is as reliable as you need it to be.

Ryan Atkins
CEO, Supercycle

Supercycle - Status Page

All systems operational

Recent notices