Gitbook 장애 기록

A faulty SQL migration caused a corrupted deployment. Our team is working on a fix.

A fix is being deployed, we are monitoring the change.

The issue affects some sites, resulting in Internal Server Error when visiting a site. We have identified a cause and expect to have a fix in near future. Not all sites are affected due to caching.

As part of the reorderable page actions last week we migrated `pageActions` into a new shape and removed old values. Our published content wasn't properly updated to no longer check the old values, and thus errored out because it was checking data that wasn't there (depending on what version of the customization object you were served, some sites were still using the old shape and thus functioned). We're mitigating by properly checking both data shapes for the foreseeable future, until all sites are completely using the new format.

The fix released has solved the issue, all systems should be operational now.

We are investigating an increased error rate in our API. This only affects the GitBook app, and public content is working as expected.

A fix has been deployed and the app and the API are responding normally. We are continuing to monitor the performance.

The incident has been resolved, all systems have returned to normal.

We are seeing a decrease in our database performance causing various actions in the app to hang. Our engineering team is investigating the root cause.

A fix has been deployed, we are monitoring the performance.

The fix has improved the performance back to their normal levels.

We’re currently investigating an issue affecting some published sites on `*.gitbook.io`

We issued a fix and have monitored the results of that fix. Pages are now loading at normal rates and our logs are looking healthy.

Due to an outage in Inngest, Git Sync operations are being affected and are being delayed. We are monitoring changes closely. Inngest outage:

The Inngest incident causing Git Sync delays has been resolved. Everything should be back to normal! Please note that any syncs that didn't go through during the incident won't be retried automatically, so you'll need to push again to trigger a new sync. If you're still seeing any issues, reach out and we'll take a look!

We are currently experiencing a degradation in API performance causing various endpoints to hang.

We have identified the issue and have mitigated it. We are currently monitoring the API and working on a root-cause fix.

We have issued the root-cause fix and are monitoring the results.

Indicators are pointing to a healthy state now after the fix. Everything should be operating well.

Between 14:36 and 14:43 UTC, all GitBook documentation sites using a custom domain were returning a 404 "Not Found" error, making documentation unavailable to users. The issue was caused by a problem introduced by an upstream provider that affected our deployment. The incident was resolved by rolling back the deployment, after which all sites were confirmed to be fully operational.

GitBook is experiencing a downtime site wide. Both the GitBook app and published documentation are affected. Our team is working to revert the situation at this time.

A fix has been deployed, and we are observing the situation going back to normal.

Both the application and published content have recovered fully after the fix was released and should be working as expected.

We are aware of issues loading public content and the GitBook app. The issue is related to a Cloudflare outage. More information on the incident can be found [here](https://www.cloudflarestatus.com/incidents/lfrm31y6sw9q).

Cloudflare has implemented a [fix for the issue](https://www.cloudflarestatus.com/incidents/lfrm31y6sw9q). Public content and the GitBook app are back.

[Cloudflare have resolved](https://www.cloudflarestatus.com/incidents/lfrm31y6sw9q) the incident and we have seen full recovery of public content and the app.

We've identified an ongoing incident with Cloudflare, who have updated their status page [here](https://www.cloudflarestatus.com/incidents/8gmgl950y3h7). This is impacting change requests, where creating new requests or merging existing ones is working intermittently.

We are still monitoring the current incident. Cloudflare has identified and is resolving the issue, and has updated its status page [here](https://www.cloudflarestatus.com/incidents/8gmgl950y3h7).

Cloudflare has [confirmed](https://www.cloudflarestatus.com/incidents/8gmgl950y3h7) that the incident has now been resolved, and we are seeing that change requests are working as expected.

We are aware of issues in loading public content in GitBook. Our team has identified the issue, and it's related to an issue in AWS, currently affecting several upstream providers. More details on the AWS outage here: https://health.aws.amazon.com/health/status Our team is looking into mitigation options.

We are seeing improvements in loading times and performance overall following updates on AWS' outage. We continue to monitor the performance.

GitBook is back to normal and fully functional.

This means that you cannot use AI search, GitBook Assistant or any other AI based features at this moment. Our team is investigating.

We have issued a temporary fix, and the AI features should be working as expected. Our team is still investigating to fix the underlying issue.

The issue was identified and fixed. All AI features are working as expected.