Admin Portal Access Issues and Post-Purchase Offer Rendering Errors

We are currently investigating an issue with users unable to access the admin console. All on-site functionality remains functional.

We are continuing to investigate this issue.

We're continuing to investigate the admin portal access issues. We've also identified that post-purchase offers are not rendering correctly for some shops. Our team is actively investigating this as well and will share updates as we learn more.

The issue has been identified and a fix is being implemented.

A fix has been implemented. We're now seeing admin access restored and post-purchase offers rendering correctly on the impacted sites we're aware of. We are monitoring to ensure stability at this time.

All tests are showing positive results and we're confirming these issues as resolved at this time. We'll continue investigating root cause and will follow up with a post mortem once that's complete.

What happened:

On July 24, between approximately 2:35 AM and 4:28 AM ET, our admin portal was briefly unavailable, and Post-Purchase Offers were not rendering for merchants using that feature. This occurred overnight during routine system maintenance.

Your storefront, checkout, and all on-site widgets and recommendations continued to function normally throughout this window. This was limited to the admin portal and the post-purchase offer surface.

‌

Root cause:

As part of our regular maintenance, we periodically clean up older, unused versions of our internal software to keep things running efficiently. During one of these routine cleanups, the process referenced a slightly outdated snapshot of what was currently live, rather than the most up to date version. This led it to retain an older version instead of the current one.

Because the affected systems continued running on cached versions in the meantime, there was no immediate impact. The issue only became visible several hours later, when those systems cycled as part of normal operations.

Resolution Once identified, our team restored service by rolling back to the last known good version, resolving the issue. We then re-released the current version to fully close things out and proactively checked our other internal systems to confirm everything else was running as expected.

What we're doing to prevent this from happening again:

• Updating our cleanup process to always reference the live, running version in real time, rather than a snapshot taken earlier. We've already tested this improvement successfully.
• Adding earlier automated alerting so our team can catch and resolve this type of issue even faster in the future.
• Continuing to strengthen our internal tooling so routine maintenance like this is even more reliable going forward.