Makeswift's Incident on June 11, 2026
June 11th, 2026
Incident Report
On June 11th, 2026, Makeswift experienced a brief spike in database saturation between approximately 12:34 p.m. and 12:38 p.m. ET. During this roughly four-minute window, Makeswift's API experienced significantly elevated response times and eventual unavailability, causing slow page loads or failed requests for sites that do not implement caching, with Catalyst being the most notable example. The Makeswift Builder was also unavailable during this window.We apologize for the disruption this caused and we're taking steps to make sure this doesn't happen again.
What happened?
The saturation was triggered when a large site published changes. Due to a bug in the runtime version that the site was on, this resulted in API calls being made faster than our API was designed to consume, flooding it with a wave of requests at a volume it was not built to absorb.For an average site on Makeswift, this behavior would not cause a problem. However, this particular site is one of the largest on Makeswift and makes an expensive query that, at scale, can cause saturation.When the volume of requests exceeded what our API could absorb, our database became saturated, leading to elevated latency and, in some cases, request timeouts.The underlying bug was introduced during a refactor of the runtime, and only affects sites on an older Next.js routing system. Nearly all active hosts are on the current routing system, which is why this issue remained latent until the affected site recently upgraded its runtime.
Incident Response Timeline
All times below are in the Eastern Timezone on June 11th, 2026.
12:32 PM: A change was published to a large customer site.
12:33 PM: Host API latency began climbing significantly, backed by a spike in database saturation.
12:37 PM: Internal alerting fired, paging on-call engineers. API requests began timing out and the Builder stopped loading.
12:38 PM: API latency and database saturation returned to normal levels.
12:39 PM: Incident declared.
12:54 PM: Identified the spike as a high volume of API requests originating from the published site within a short time frame; paused further publishing.
2:22 PM: Identified the contributing bug in our runtime package.
3:43 PM: Incident declared resolved.
5:08 PM: Released a patched version of the runtime.
6:16 PM: Affected site deployed with the latest runtime version.
The API was experiencing elevated latency for approximately 4 minutes.
Short Term Resolution
- We have implemented this fix in our `@makeswift/runtime` package, and released patched versions for all affected versions (
0.25through0.28). If you're onv0.25tov0.28of the runtime, we recommend upgrading your host to their latest patch versions:0.25.5,0.26.6,0.27.4, and0.28.7.
Long Term Mitigation
- We have added regression test coverage to catch this class of bug before it ships, closing the gap in our test suite that allowed this issue through.
- To prevent incidents from impacting site traffic, we are building a CDN for Makeswift's API designed to handle high traffic volumes and thundering herds, decoupling our API uptime from the data sites depend on for page visits. This project was already on our roadmap for this summer, and this incident underscores its importance.
- For Catalyst, we are continuing to work on ensuring page view traffic does not pass through to Makeswift's API, so that incidents of this type are significantly less likely to impact sites.
We apologize for the impact this had and appreciate your patience. If you have any questions, reach out to us at support@makeswift.com.