Spreedly 장애 기록

1. 2026년 9월 2일 PM 3:11 UTCinvestigating

   Spreedly는 Transactional APIs에서 잠재적으로 결과가 발생했습니다. 우리는 가능한 한 빨리 경고하기 위해 노력의 초기 알림을 제공하지만, 우리는 여전히 실제 범위와 영향을 조사하고 더 많은 세부 사항이 가능한 한 빨리 업데이트를 제공 할 것입니다. 감사합니다.

2. 2026년 9월 2일 PM 3:57 UTCmonitoring

   우리는 최근 데이터베이스 유지 보수와 관련있는 일부 지불 처리 (특히 구매, 캡처 및 지불 방법 생성)에 영향을 미치는 성능 지연을 해결하기 위해 노력하고 있습니다. 수정은 언웨이이며, 우리 팀은 적극적으로 시스템을 모니터링하고 있습니다. 실시간 업데이트에 대해 다시 확인하시기 바랍니다.

3. 2026년 9월 2일 PM 9:32 UTCresolved

   이 사건은 해결되었습니다.

공식 인시던트 업데이트를 자동 번역했습니다.

1. 2026년 8월 14일 PM 4:14 UTCinvestigating

   We are currently investigating an increase in 5xx errors on the Core Transactional API. Some transactions may be failing or returning unexpected error responses during this time. We will provide an update as soon as we have more information.

2. 2026년 8월 14일 PM 4:34 UTCresolved

   A fix has been implemented and error rates have returned to normal. Some transactions may have failed or returned errors during the affected window. We're continuing to monitor and will follow up with a more detailed post-incident summary.

3. 2026년 9월 3일 PM 8:55 UTCpostmortem

   ### Service Disruption on Transaction Endpoints ### ‌High Level Summary On August 14, 2026, Spreedly experienced a brief service disruption affecting some transaction-related API endpoints for approximately 20 minutes \(4:06 PM – 4:26 PM ET\). During this window, some requests to endpoints involving transaction lookups, captures, voids, credits, and payment method operations may have received error responses. The issue was identified and resolved quickly by our engineering team. ### What Happened A routine database schema update caused a subset of existing database connections to reference outdated schema information, resulting in intermittent errors on affected requests. Once identified, our team refreshed the affected database connections, which fully resolved the issue. ### Next Steps We are reviewing our schema migration process to add additional safeguards that will prevent this class of issue from recurring, and are improving automated testing coverage for this scenario. We have added auto-retry and recovery mechanisms so that if a connection does reference an outdated schema it will not affect transactions going forward.

1. 2026년 7월 27일 PM 1:50 UTCinvestigating

   우리는 현재이 문제를 조사하고 있습니다.

2. 2026년 7월 27일 PM 2:00 UTCidentified

   문제는 식별되고 수정이 구현됩니다.

3. 2026년 7월 27일 PM 3:00 UTCmonitoring

   Spreedly는 핵심 거래 성과에 영향을 미칠 수 있는 문제점을 검출했습니다. 이제 문제가 해결 될 것으로 나타났습니다. 우리 팀은 지속적으로 안정성을 보장하기 위해 적극적으로 모니터링 시스템입니다. 우리는 우리의 조사가 전체 범위와 충격에 완료되면 최종 업데이트를 제공 할 것입니다. 최신 업데이트에 대한 StatusPage를 계속하십시오.

4. 2026년 7월 27일 PM 9:41 UTCresolved

   밀접한 감시 및 확인 후 모든 시스템은 안정적이고 기능적인 것으로 예상됩니다. 이 사건은 해결되었습니다. 더 많은 고객 영향이 예상되지 않습니다. 우리는 사건과 잔여 충격의 원인에 관한 우리의 조사를 완료하고 있습니다. 우리는 어떤 inconvenience 또는 혼란든지를 위해 apologize.

5. 2026년 9월 3일 PM 8:57 UTCpostmortem

   ## 높은 수준의 요약 7월 27, 2026일, 구매, 승인 및 webhooks를 포함한 거래 요청의 하위 집합에 대한 Spreedly 경험있는 간헐적 오류. ### 신청 데이터베이스 마이그레이션은 저장된 지불 방법을 관리하는 내부 서비스의 schema를 업데이트. 기존의 연결에 의한 오류는 stale 캐시를 참조합니다. 우리는 변경을 압연하고 그 연결을 순환, 완전히 문제를 해결. ## 다음 단계 이 문서는 저작권법의 보호를 받습니다. 이 문서는 저작권법의 보호를 받습니다. 우리는 자동 복원 및 복구 메커니즘을 추가했습니다. 연결이 outdated schema를 참조하면 거래에 영향을 미치지 않습니다.

공식 인시던트 업데이트를 자동 번역했습니다.

1. 2026년 7월 11일 PM 2:33 UTCinvestigating

   특정 API 요청에 영향을 미치는 높은 오류율을 조사하고 있습니다. 일부 목록과 retrieval endpoints는 우리가 뿌리 원인을 확인하기 위해 작동하면서 간헐적 인 500 응답을 반환 할 수 있습니다.

2. 2026년 7월 11일 PM 3:38 UTCresolved

   우리는 특정 목록과 retrieval endpoints에 대한 간헐적인 500 응답을 일으키는 문제를 확인하고 해결했습니다. 구성 업데이트는 조기 종료 된 요청을 방지하기 위해 배포되었으며, 우리는 서비스를 계속 모니터링하고 있습니다.

공식 인시던트 업데이트를 자동 번역했습니다.

1. 2026년 7월 8일 PM 3:22 UTCinvestigating

   Spreedly는 Cybersource Credit 거래에서 잠재적으로 결과가 발생했습니다. 우리는 가능한 한 빨리 경고하기 위해이 초기 알림을 제공하지만, 우리는 여전히 실제 범위와 영향을 조사하고 있습니다. 우리는 더 많은 세부 사항이 가능한 한 빨리 업데이트를 제공 할 것입니다.

2. 2026년 7월 8일 PM 4:08 UTCidentified

   우리는 15 : 00 및 15 : 05 UTC 사이에 발생 한 일시적 내부 DNS 해상도 오류가 Spreedly 서비스에서 500 오류를 식별했습니다. 우리는 전체 충격을 조사하고 추가 정보를 제공 할 것입니다.

3. 2026년 7월 8일 PM 4:54 UTCresolved

   밀접한 감시 및 확인 후 모든 시스템은 안정적이고 기능적인 것으로 예상됩니다. 이 사건은 해결되었습니다. 더 많은 고객 영향이 예상되지 않습니다. 우리는 사건과 잔여 충격의 원인에 관한 우리의 조사를 완료하고 있습니다. 우리는 어떤 inconvenience 또는 혼란든지를 위해 apologize.

공식 인시던트 업데이트를 자동 번역했습니다.

1. 2026년 6월 17일 PM 2:55 UTCinvestigating

   Spreedly는 잠재적으로 비동기 거래와 차별화된 성능에 대한 문제점을 발견했습니다. 우리는 가능한 한 빨리 경고하기 위해 노력의 초기 알림을 제공하지만, 우리는 여전히 실제 범위와 영향을 조사하고 더 많은 세부 사항이 가능한 한 빨리 업데이트를 제공 할 것입니다. 감사합니다.

2. 2026년 6월 17일 PM 3:00 UTCmonitoring

   우리는 비동기 거래를 해결하고 안정화했습니다. 우리의 팀은 상황을 밀접하게 감시하고 추가적인 충격이 없습니다. 최신 업데이트에 대한 StatusPage를 계속하십시오.

3. 2026년 6월 17일 PM 4:03 UTCmonitoring

   우리는 더 많은 문제 또는 추가적인 충격을 위해 밀접하게 감시하고 있습니다. 최신 업데이트에 대한 StatusPage를 계속하십시오.

4. 2026년 6월 17일 PM 9:42 UTCresolved

   감시의 기간 후에, 우리는 문제점이 해결되고 체계가 정상적으로 작동된다는 것을 확인했습니다. 문제가 계속되면 지원해 주세요.

5. 2026년 6월 24일 PM 8:10 UTCpostmortem

   6월 24, 2026 — 비동기 거래 처리 중단 **높은 수준의 요약** 6월 17, 2026에서 약 10:02 AM에서 10:41 AM UTC, Spreedly는 비동기 게이트웨이 거래와 3DS 인증 흐름에 영향을 미치는 혼란을 경험했습니다. 약 1,634건의 거래는 확인된 결과가 없는 주에 남아있으며, 그 기간 동안 총 거래량의 1% 미만을 나타냅니다. **Happened ** 코드 배포는 새로 배포 및 이전 실행 애플리케이션 인스턴스 간의 호환성 문제를 도입했습니다. 롤백 프로세스 중 일부 요청은 혼합 된 코드 버전에서 경로를 거쳐 비동기 거래 상태 업데이트를 처리 할 때 실패를 발생시킵니다. Affected 거래 유형에는 비동기 게이트웨이 거래 \ (게이트웨이 특정 3DS 및 기타 비동기 검증, 승인 및 구매) 및 3DS 글로벌 인증을 포함합니다. 이 창에서 영향을받은 흐름을 입력하는 거래는 여러분의 게이트웨이의 거래 상태는 Spreedly 보고서와 다를 수 있습니다. **다음 단계** 우리는 6월 17일 오전 10:02–10:41 UTC 창에서 트랜잭션을 검토하는 것이 좋습니다. 검증은 안전하게 회복될 수 있습니다. 권한은 재발송하기 전에 게이트웨이에 취약해야합니다. 구매는 복제의 위험을 수행, 우리는 재발송하기 전에 환불을 권장합니다. Spreedly support에 도달하면 영향을받는 거래 식별 또는 적절한 치료 경로를 결정해야합니다. 우리는 개선 된 배포를 통해이 종류의 문제를 방지하기 위해 적극적으로 노력하고 있으며 자동화 된 롤백 탐지. -Spreedly 팀

공식 인시던트 업데이트를 자동 번역했습니다.

1. 2026년 5월 8일 PM 6:00 UTCmonitoring

   A fix has been implemented and we are monitoring results.

2. 2026년 5월 8일 PM 6:05 UTCresolved

   This incident has been resolved.

1. 2026년 3월 19일 PM 3:13 UTCidentified

   We are currently investigating performance of the Dodgeball service.

2. 2026년 3월 19일 PM 3:30 UTCidentified

   We are working to resolve an issue with delayed processing of requests to the Dodgeball services.

3. 2026년 3월 19일 PM 3:55 UTCmonitoring

   We’ve implemented a fix and stabilized the Dodgeball component that was experiencing a partial outage. We are continuing to monitor

4. 2026년 3월 19일 PM 4:07 UTCmonitoring

   After closely monitoring and confirming that all systems are stabilized and functioning as expected, this incident is considered resolved. No further customer impact is expected. We apologize for any inconvenience or disruption.

5. 2026년 3월 19일 PM 4:09 UTCresolved

   This incident has been resolved.

1. 2026년 3월 16일 PM 5:45 UTCinvestigating

   We are currently investigating elevated errors on API endpoints

2. 2026년 3월 16일 PM 5:51 UTCmonitoring

   A fix has been implemented and we are monitoring to ensure services are operating correctly

3. 2026년 3월 16일 PM 6:07 UTCresolved

   After closely monitoring and confirming that all systems are stabilized and functioning as expected, this incident is considered resolved. No further customer impact is expected. We are completing our investigation concerning the causes of the incident and any residual impact. We apologize for any inconvenience or disruption.

4. 2026년 3월 19일 PM 7:07 UTCpostmortem

   ### Mar 18th 2026 — _Decryption Errors on API endpoints_ ### Summary Between 5:40 PM and 5:43 PM UTC on March 16, 2026, we experienced a brief issue during a key rotation that caused a small subset of API requests to fail due to decryption errors. These errors occurred within our system before any external processing took place. The issue was quickly identified, and traffic was immediately redirected to a stable environment, restoring normal service. ### What Happened At 5:36 PM UTC, we routed traffic to a newly deployed cluster of our encryption service running with the latest security keys. Shortly after, we observed decryption errors affecting a small subset of records, which resulted in some requests returning 500 error responses. We promptly redirected traffic back to the existing cluster, restoring normal operation. Further investigation determined that these records were still associated with a key in the process of rotation, and we updated them as a follow-up action to prevent the issue from recurring. ### Next Steps We have enhanced our process to ensure that a key’s usage is fully verified before it is removed from the encryption service. -The Spreedly Team

1. 2026년 2월 26일 PM 4:45 UTCinvestigating

   We are currently investigating an issue affecting customers using Dodgeball. The production database backing Dodgeball has reached critically low available storage, which is expected to result in elevated error rates. Customers may begin experiencing errors within the next 10 minutes as the system’s ability to process requests degrades. Our engineering team is actively investigating and working to restore normal capacity as quickly as possible. We will provide further updates as more information becomes available.

2. 2026년 2월 26일 PM 4:59 UTCmonitoring

   The engineering team has successfully resolved the storage issue affecting the Dodgeball production PostgreSQL instance. Capacity has been restored and the system is now operating normally. At this time, we are not observing any elevated error rates for customers using Dodgeball. We will continue to monitor the system closely to ensure stability. Further updates will be provided if necessary.

3. 2026년 2월 26일 PM 6:44 UTCresolved

   The storage issue affecting the Dodgeball production instance has been fully resolved. Capacity has been restored and systems are operating normally. We are no longer observing elevated error rates for customers using Dodgeball. Thank you for your patience while we worked to resolve this issue.

1. 2026년 2월 25일 PM 7:18 UTCinvestigating

   We are currently investigating performance of the Dodgeball service.

2. 2026년 2월 25일 PM 7:48 UTCidentified

   We are working to resolve an issue with processing of Checkpoints that is causing latency and errors for requests to the Dodgeball services.

3. 2026년 2월 25일 PM 8:35 UTCmonitoring

   Service has resumed and we are monitoring for further impact. We apologize for any inconvenience this has caused

4. 2026년 2월 25일 PM 9:52 UTCresolved

   After closely monitoring and confirming that all systems are stabilized and functioning as expected, this incident is considered resolved. No further customer impact is expected. We are completing our investigation concerning the causes of the incident and any residual impact. We apologize for any inconvenience or disruption.

1. 2026년 2월 5일 PM 3:56 UTCinvestigating

   We are currently investigating an issue where credit transactions referencing Capture transactions are failing in production. Some customers may see errors when attempting to issue credits against Capture transactions. Capture transactions themselves are processing successfully. Our engineering team is actively investigating this behavior to determine the cause and confirm whether this is a regression or an unintended change. We will provide updates as soon as more information is available.

2. 2026년 2월 5일 PM 4:25 UTCmonitoring

   We’ve deployed a revert to address the issue where credit transactions referencing Capture transactions were failing. We’re actively monitoring to ensure credit requests referencing Captures are processing as expected. At this time, we’re not seeing further errors related to this behavior.

3. 2026년 2월 5일 PM 5:01 UTCresolved

   After monitoring the system post-deploy, we’re no longer seeing errors related to credits referencing Capture transactions, and processing has returned to normal. Thank you for your patience.

4. 2026년 2월 10일 PM 5:22 UTCpostmortem

   ### February 10, 2026 — Credits Referencing Capture Transactions Failing ### ‌High Level Summary Between approximately 1:56 PM UTC and 4:15 PM UTC on February 5, 2026, credit transactions where the reference transaction was a capture failed with the error "The specified `reference_token` corresponds to a transaction which cannot be used as a reference." ### What Happened At approximately 1:56 PM UTC on February 5, 2026, Spreedly rolled out a change to support credits on the CyberSource Rest gateway. This included a change in the way we validated reference transaction types used in credits. Specifically, there was an unintended side-effect which caused captures to no longer be treated as a valid reference type for credits on gateways other than CyberSource Rest. Once this was discovered, the code was reverted at 4:01PM UTC. By 4:15 PM UTC, credit requests against captures were fully functional. ### Next Steps * Any credit attempts that failed with this error are safe to retry * We are making improvements to our code review processes and regression test suite to ensure we catch similar errors before they make it into production * We have identified a gap in our observability surrounding validation errors that we will be prioritizing ‌ -The Spreedly Team

1. 2026년 2월 4일 PM 6:15 UTCinvestigating

   Spreedly has detected an issue that is affecting multiple customers using Payflow Pro where some transactions fail with the error: “Invalid Request. Malicious string found in XML document”. While we are providing an early notification in an effort to alert you as quickly as possible, we are still investigating the actual scope and impact and will provide an update as soon as more details are available. Thank you for your patience.

2. 2026년 2월 4일 PM 7:52 UTCmonitoring

   We are no longer seeing errors at this time, and the issue appears to have subsided. We are working with Payflow Pro to identify the root cause. Our team continues to monitor the situation closely to ensure stability. Please continue to follow our StatusPage for the latest updates.

3. 2026년 2월 4일 PM 7:56 UTCresolved

   This incident has been resolved.

1. 2026년 1월 27일 PM 7:59 UTCinvestigating

   Spreedly has detected an issue related to a dodgeball component experiencing a partial outage. While we are providing an early notification in an effort to alert you as quickly as possible, we are still investigating the actual scope and impact and will provide an update as soon as more details are available. Thank you for your patience.

2. 2026년 1월 27일 PM 8:12 UTCresolved

   We’ve implemented a fix and stabilized the Dodgeball component that was experiencing a partial outage. After closely monitoring and confirming that all systems are stabilized and functioning as expected, this incident is considered resolved. No further customer impact is expected. We apologize for any inconvenience or disruption.

1. 2026년 1월 27일 PM 7:08 UTCinvestigating

   Spreedly has detected an issue potentially resulting in a high number of pending 3DS2 Global transactions. While we are providing an early notification in an effort to alert you as quickly as possible, we are still investigating the actual scope and impact and will provide an update as soon as more details are available. Thank you for your patience.

2. 2026년 1월 27일 PM 7:25 UTCmonitoring

   We’ve implemented a fix and stabilized the 3DS2 Global transactions. Our team is monitoring the situation closely to ensure there are no additional impacts. Please continue to follow our StatusPage for the latest update.

3. 2026년 1월 27일 PM 7:29 UTCresolved

   After closely monitoring and confirming that all systems are stabilized and functioning as expected, this incident is considered resolved. No further customer impact is expected. We are completing our investigation concerning the causes of the incident and any residual impact. We apologize for any inconvenience or disruption.

4. 2026년 2월 3일 PM 3:38 UTCpostmortem

   ### February 3rd, 2026 — 3DS2 Global Transactions Pending ‌ Between 5:40 pm UTC and 7:24 PM UTC on January 27, 2026 the latest version of the Spreedly iFrame \(`iframe-v1`\) was configured to use version 1.180 which contained a bug related to the handling of device fingerprinting for `Sca::Authentications` in an end user's browser. As a result there was an elevated number of `Sca::Authentications` stuck in a pending flow with a `required_action` of `device_fingerprint`. These transactions were held in a pending state and never completed. Our team identified the issue and updated the version of iFrame \(`iframe-v1`\) to utilize `1.179`. ### What Happened At 5:40 pm UTC on January 27th, 2026 a new version of iFrame was released \(1.180\) and updated the version to be used in the `iframe-v1.min.js` asset to point to `1.180`. This change was meant to improve the error handling for the loading of 3rd party critical SDKs however it caused the SDK to enter a permanently error-ed state when attempting to perform device fingerprinting due to a race condition with loading the SDKs. At 7:20 pm UTC on January 27th, 2026 the change was rolled back and by 7:24 pm UTC, 2026 the `iframe-v1.min.js` asset globally was pointing to version `1.179`. ### Next Steps We have identified a gap in our observability and alerting within the iFrame service related to the managing of 3DS lifecycle flows that we will be prioritizing. -The Spreedly Team

1. 2026년 1월 9일 PM 6:46 UTCinvestigating

   We are currently investigating this issue.

2. 2026년 1월 9일 PM 6:52 UTCinvestigating

   We are continuing to investigate this issue.

3. 2026년 1월 9일 PM 7:07 UTCmonitoring

   A fix has been deployed and we are seeing systems recover. We will continue to monitor to ensure stability and will provide further updates as needed.

4. 2026년 1월 9일 PM 7:10 UTCresolved

   After a period of monitoring, we have confirmed that the issue has been resolved and systems are operating normally. If you continue to experience any issues, please reach out to Support.

1. 2025년 12월 29일 PM 4:50 UTCresolved

   Between 7:00 PM EST on Saturday, December 27th, and 1:00 AM EST on Sunday, December 28th, we identified a period of higher-than-normal latency caused by a pre-planned customer event that resulted in a significant traffic spike. Although we did not observe any service failures or timeouts, our systems experienced degraded service due to the increased latency. Systems recovered quickly once peak traffic subsided and are now operating normally. We expect performance to remain stable moving forward.

1. 2025년 12월 16일 PM 9:56 UTCresolved

   Between 3:59 and 4:03 EST, customers may have experienced degraded performance when processing transactions that required 3D Secure (3DS) authentication via our Global 3DS service. Performance has since returned to normal, and no further impact has been observed. If you have any questions about potential impact to your transactions during this time window, please reach out to Spreedly Support for assistance.

2. 2025년 12월 18일 PM 9:26 UTCpostmortem

   # _December 18, 2025 — Delayed Processing of Some 3DS Global Transactions_ ‌ _A targeted update intended for Gateway-Specific 3DS processing impacted a subset of 3DS Global transactions. This caused a small number of transactions to remain in a "Pending" state rather than progressing to completion._ ### What Happened Between 08:59 PM and 09:03 PM UTC on December 16, 2025, a deployment intended to update Gateway-Specific 3DS logic inadvertently affected the processing of 3DS Global transactions. As a result, 52 transactions failed to reach a final state and remained "pending." During this 4-minute window, customers attempting a 3DS-authenticated operation would have encountered an interruption, preventing their transactions from completing. Our team identified the impact immediately and rolled back the change to restore normal service. ### Next Steps _We have successfully identified and corrected the underlying logic error within the change. We are also enhancing our automated deployment tests to better catch cross-module dependencies of this nature. We sincerely apologize for any disruption this caused to your operations._ ‌ — The Spreedly Team

1. 2025년 12월 9일 PM 3:58 UTCinvestigating

   We are investigating an issue where roughly 30–40% of transactions are not appearing in Spreedly reporting since 12/03. Primary datastores and secondary services show no data loss, indicating the issue lies within the reporting ingestion process. Please reach out to Spreedly Support if you need assistance.

2. 2025년 12월 10일 AM 6:53 UTCinvestigating

   Our team continues to investigate the root cause of the missing transaction data appearing in Spreedly reporting. All primary datastores and transactional processing are functioning normally, and there is no indication of data loss at the source. The issue appears to be isolated to the reporting ingestion process.

3. 2025년 12월 10일 PM 2:42 UTCmonitoring

   We have identified and fixed the issue that was causing some transaction data to be missing from Spreedly reporting. Our team is now working on backfilling the missing reporting data to ensure completeness and accuracy. All primary processing systems continue to operate normally, and there is no impact to transaction processing.

4. 2025년 12월 10일 PM 8:04 UTCresolved

   Our team has completed the backfill process, and all missing transaction data has now been fully restored in Spreedly reporting. Transaction reporting is up to date and functioning as expected. All primary processing systems continued to operate normally throughout this incident, and no transaction processing impact occurred.

5. 2026년 1월 2일 PM 9:44 UTCpostmortem

   ### _Postmortem: January 2, 2026_ — _Transaction Reporting Degraded_ ### ‌High Level Summary Select data related to transactions did not make it into our downstream systems, causing inaccurate reports in [app.spreedly.com](http://app.spreedly.com) ### What Happened After completing system updates on December 3, 2025, approximately 30-40% of data related to transactions did not make it into our reporting tools due to a latent bug in one of our supporting applications. This caused inaccurate reports in [app.spreedly.com](http://app.spreedly.com). This did not impact any primary system functionality and no data was lost. ### Next Steps * Our team has since identified and fixed the affected application. The missing data was backfilled and subsequent data continues to flow normally * We are adding additional monitoring in order to more quickly flag discrepancies * We are continuing to make updates to simplify our data pipeline that will prevent similar issues in the future -The Spreedly Team

1. 2025년 10월 20일 PM 3:41 UTCidentified

   The major AWS outage on October 20, 2025, has been largely resolved for most affected services, though some residual issues may persist. Spreedly’s core platform remains stable and fully available. However, due to downstream impact from the AWS outage, some of Spreedly’s indexing endpoints and callback processing are experiencing delays. This is related to dependencies on third-party systems that were affected by the outage and are still undergoing recovery. As a result, you may continue to see intermittent latency or delayed responses for API requests involving these systems. We are actively monitoring the situation as AWS and its downstream services continue recovery efforts. For ongoing updates, please refer to the AWS status page here

2. 2025년 10월 20일 PM 8:22 UTCresolved

   After closely monitoring and confirming that all systems are stabilized and functioning as expected, this incident is considered resolved. No further customer impact is expected.

3. 2025년 10월 23일 PM 6:27 UTCpostmortem

   ### Postmortem: October 24, 2025 — _AWS outage_ Monday's AWS outage had downstream impact on some 3rd-party systems utilized by Spreedly's platform. Namely, our indexing endpoints and callbacks lagged behind due to the outage and recovery steps. ### What Happened Spreedly's 3DS global service and legacy smart routing product both faced a period of failures due to 3rd-party systems relying on AWS. ### Next Steps Spreedly will improve its response to our alerts platform, and are actively working on multi-region support slated for 2026, which should mitigate the impact of similar events going forward. -The Spreedly Team