16 Cdnjs incidents · 2018年12月 — official updates, affected components, duration and resolution details.
我们知道一个影响图书馆搜索能力的问题。 依赖api.cdnjs.com/library endpoint,浏览cdnjs.com,或整合cdnjs的第三方服务的用户在搜索库时可能会收到错误. 在CDN上获取资源不受影响. 我们已经找出一个问题 我们的上游搜索提供者 并且已经把问题升级到他们.
我们的上游搜索提供者已经实施了一个固定,现在又可以使用cdnjs上的搜索图书馆. 我们仍在与我们的搜索提供商合作,以了解这个问题的原因,并确保这不会再次影响服务.
我们确认Cdnjs上的搜索图书馆仍然可用。 我们的上游搜索提供商也确认了他们的结局.
自动翻译自官方事件更新。
We're investigating an outage with the cdnjs API. We believe this to be caused by an upstream provider issue.
cdnjs' API is impacted by an ongoing outage with Cloudflare services: https://www.cloudflarestatus.com/incidents/25r9t0vz99rp
Cloudflare has indicated that it is beginning to see recovery of its services. We continue to monitor the impact to cdnjs' API and will provide an update when it has recovered.
At this time, we're seeing cdnjs' API service recover and return to normal operation. We'll continue to monitor the service until the upstream provider indicates resolution and we're happy the service is stable.
Cloudflare has now marked its incident as resolved, and we are confident that the cdnjs API is now operating as normal.
We are aware of an issue impacting multiple libraries where SRI hashes have changed. We have found the cause of the change to the file contents/hashes and are actively working to restore the previous contents/hashes. No files on the CDN have been compromised, the contents/hashes have changed due to libraries being reprocessed with different minification logic. We are tracking this on GitHub in https://github.com/cdnjs/cdnjs/issues/14210
A backfill has been completed that is believed to have restored files to their previous states, with the SRI hashes reverted to what they were prior to this incident. We'll continue to monitor for reports of any hash mismatches, which should be reported in the issue on GitHub at https://github.com/cdnjs/cdnjs/issues/14210
At this time we believe the fix to restore file content and SRI hashes is working as expected.
We are aware that the cdnjs API is returning 502 responses for many requests. We are investigating this issue. (Edit: This incident has been back-dated to the believed initial starting point of impact for requests. Incident was initially declared at 11:30 UTC.)
We are continuing to investigate the 502 issue with the API. It appears to be impacting a limited subset of requests at this time.
At this time the API is now returning valid responses for all requests. We will continue to monitor the situation to ensure this remains true.
This incident has been resolved.
We are aware that some requests to the API may receive invalid responses. We are looking into this.
Our team are online and have temporarily restored the API service. We continue to investigate the root cause and look for a solution.
We are confident that a fix is now in place for the intermittent issues previously experienced by some users with the API service. We will continue to monitor the situation to ensure this remains the case.
The API service has now been restored.
We are aware that the cdnjs API is currently returning a 502 Bad Gateway error for requests. We are investigating this and will update once we know more.
We continue to investigate the 502 error issue with the API and are aware that this may also be impacting accessing library information on the cdnjs website as this is backed by the cdnjs API service.
We have implemented a fix for the 502 issues -- the API service, as well as the website, appears to be operational once again. We will continue to monitor the situation to ensure this remains the case.
We are aware of a secondary issue where some requests are receiving a 530/1016 Origin DNS error and we are working to resolve this.
A further fix has been implemented and the API service appears to be healthy now. We will continue to monitor this.
We are confident that the API service and the website are both stable.
We are aware that the cdnjs API is currently returning an Argo Tunnel error. We are investigating this and will provide an update as soon as we know more.
The API is now responding nominally. We will continue to monitor the situation to ensure the cdnjs API remains stable.
This incident has been resolved.
We are aware that requests to cdnjs.cloudflare.com are currently returning a bad SSL response. We are working to establish the root cause and get a fix implemented.
The team at Cloudflare are aware of the issue and are working to fix it. You can track their updates at https://www.cloudflarestatus.com/incidents/vtlmfzqmc8h7 -- we will continue to share updates here as well.
Cloudflare has now implemented a fix for the CDN and SSL errors should begin to resolve. Users in mainland China may still see issues connecting to cdnjs and we continue to work to resolve this.
A fix has been implemented that should allow users in mainland China to access cdnjs. Cloudflare will continue to monitor the situation.
At this time we are confident that access to the CDN has been restored for all users of cdnjs.
We are aware of reports that some libraries on the cdnjs website and API are resulting in an error response. We have notified our team and are looking into the root cause of this issue.
We have identified the source of the failure, the metadata used for the website and API for packages data was unable to generate correctly, resulting in the data used being corrupted. We are working to mitigate this and regenerate the data.
The cdnjs API is now correctly responding to library requests, however, we are still working to resolve the issue present on the website.
The cdnjs website and API are now both using updated packages metadata, allowing them to correctly respond to requests for libraries without error. We will continue to monitor both the API and website to ensure this doesn't happen again.
We are confident at this time that we have resolved the errors seen with the website & API, both are now handling library requests correctly.
We are aware of an issue affecting requests for certain assets resulting in a 522 or 523 response. We are currently investigating the source of this issue.
We believe this is due to the ongoing incident Cloudflare has: https://cfl.re/2VbxQ96 - Requests to cached assets should work but any request for an asset not cached on cdnjs is likely to fail at present.
It would appear the origins for cdnjs.cloudflare.com are now available again and requests to uncached cdnjs assets should now work as expected. We will continue to monitor for a while to ensure this is stable. For more information on the larger Cloudflare outage, please follow their status page updates: https://cfl.re/2VbxQ96
We are confident at this time that all cdnjs assets are now available on the CDN once again. Cloudflare is continuing to see good recovery on their incident as well: https://cfl.re/2VbxQ96
We have been made aware of an issue with the cdnjs website that means certain library pages are not showing, with an error message "version not found!" shown instead. Further, it appears that on the same libraries, their API responses do not contain the correct versions. This does not affect the libraries available on the CDN, cdnjs.cloudflare.com.
We have identified the issue as a fault within our package generation script that feeds cdnjs repository information to our website & API. We are tracking this issue on GitHub: https://github.com/cdnjs/cdnjs/issues/13654
We have now implemented a fix and believe that packages should now be showing correctly on the cdnjs website and via the API. We will continue to monitor this whilst the caches clear to ensure this issue is no longer present.
We are now confident the fix is working as expected. The cdnjs website and API are now showing libraries and their versions correctly.
We are aware of reports from users that cdnjs (website & CDN) is returning 502 Bad Gateway errors. We believe this is a larger issue with Cloudflare itself as many other websites behind them are also down. We will keep this updated as we know more.
Cloudflare have now raised an incident on their status page due to the outage: https://www.cloudflarestatus.com/incidents/tx4pgxs6zxdr
Cloudflare is stating that they have implemented a fix for this issue, we will continue to monitor our own user reports of any further outages and also monitor the Cloudflare incident.
Cloudflare is confident at this time that services have been restored to normal and users should no longer be affected by this outage. At this time we will, therefore, mark the incident as resolved. If you are still seeing any issues with the cdnjs services, please let us know through Twitter and we can look into it further.
We have been made aware of some users who are having connectivity issues to both our main website and the CDN service. We believe at this time that this is due to network issues that Cloudflare are experiencing. This is likely to also impact our API, though we have no reports of this yet. You can track the Cloudflare incident at https://www.cloudflarestatus.com/incidents/46z55mdhg0t5
Cloudflare have identified the network issue as a route leak and have their network operations team engaged in restoring services. Keep updated via their incident at https://www.cloudflarestatus.com/incidents/46z55mdhg0t5
Cloudflare are still working to resolve the route leak that is causing major outages across the internet. Follow their status updates at https://www.cloudflarestatus.com/incidents/46z55mdhg0t5
Cloudflare and the responsible network for the route leak have now fixed the issue. Connectivity should now be recovering. We will continue to monitor the effects on the cdnjs services and follow updates from Cloudflare.
Cloudflare have confirmed this issue is now resolved and all cdnjs services should be accessible again.
We have been made aware of some responses lacking CORS headers from the CDN service. We are working with Cloudflare, who have recently deployed changes, to resolve this.
Cloudflare have identified the source of the issue and are currently working to roll out a fix to the missing CORS headers.
Cloudflare have released a fix for the issue. Please let us know via Twitter or our main GitHub repository if you continue to see issues. Twitter: https://twitter.com/cdnjs or https://twitter.com/cdnjsStatus GitHub: https://github.com/cdnjs/cdnjs/issues/13324
Cloudflare have purged all affected resources from the cache and are confident that this issue has now been resolved.
It appears that the CDN is failing to serve a small minority of requests again. We have alerted Cloudflare and are waiting for their response.
To assist in the diagnosis of the issue, please post any failed requests with full request/response headers on our GitHub mega-issue: https://github.com/cdnjs/cdnjs/issues/13165
Cloudflare have identified the issue a recurrence of the CDN failure we experienced in December. They have now fixed the issue and we are monitoring the CDN to ensure service is restored.
Further reports have been made of continued issues with the CDN and these have been passed on to Cloudflare to ensure that the issue is resolved in a timely manner. We apologise for the drop in service to those affected.
This issue had resurfaced due to an automatic process with Cloudflare reinstating a broken node into the load balancer. This has now been stopped and we will continue to monitor to ensure the CDN is performing normally.
We believe that the CDN is now back to normal and service should be fully restored to cdnjs! Apologies from all of us here at cdnjs for the delay in getting this resolved and for the drop in service. Have a great weekend making use of cdnjs in all your projects!
We have been made aware that our CDN is in a small number of requests giving an invalid response that is a 404 with missing CORS headers. We are currently investigating this and will post an update here shortly.
To help us with investigating this issue we have created a GitHub mega-issue. Please post any failed requests will full request/response headers and the resource URL that failed: https://github.com/cdnjs/cdnjs/issues/13165
We believe we have identified the issue relating to the minority of requests that are receiving a bad response. We have alerted Cloudflare to this and are currently awaiting their response to resolve this.
We are aware that a higher percentage of requests are beginning to fail now. We apologise for the delay in resolving this. Please continue to post full request/response headers on the GitHub mega-issue to assist Cloudflare in resolving this: https://github.com/cdnjs/cdnjs/issues/13165
Cloudflare are now handling this issue and actively looking to resolve it as it is a failure of their infrastructure behind cdnjs that is causing this issue. They apologise for any delay due to engineering staff being away on holiday leave.
To assist Cloudflare in identifying bad origin servers for the CDN, please post any failed requests on our GitHub mega-issue with full request and response headers along with the resource URL that failed: https://github.com/cdnjs/cdnjs/issues/13165
Cloudflare have now identified the failing origin servers and they have been removed from the load balancer. CDN responses should now return to normal and cdnjs/Cloudflare will continue to monitor this.
We believe that the CDN is now serving all requests with valid responses. We sincerely apologise for the long delay in resolving this issue. We will continue to monitor the situation for a while longer to confirm all is okay.
We and Cloudflare are happy to say that the CDN is now performing normally.