Historique des pannes de Clarifai

General image recognition and embedding models had a down-time from 10:30am EST - 3:40pm EST on Monday March 16, 2026 . This is proactively identified and resolved by our Engineering team. The models are up and running now.

Due to issues on gitbook host, our documentation is not available. We are working with gitbook on resolving this issue.

While we are working with gitbook to fix the issue, we found a workaround. Go to the https://docs.clarifai.com/api-guide/api-overview/api-clients (it should be accessible) and then click 'Clarifai Guide' on the top-left corner. It should take you to the https://docs.clarifai.com/

This incident has been resolved.

We are currently investigating this issue.

Issues was identified. All services are operational now.

User cannot login to portal and SSO buttons (Google, Github) are not appearing on login screen.

We are continuing to investigate this issue.

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

This incident has been resolved.

We are currently investigating this issue.

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

We are working through a database issue that is affecting all queries.

At 6:30 am EDT, we experienced an outage on a database node as a result of an automated failover on a redundant node. This outage is affecting all customers and is currently preventing write and search operations on our cloud platform. All prediction operations continue to function normally. Our team is working diligently with our service provider to restore service. In the meantime, we will continue to keep you updated as we restore our systems. We understand the impact this is potentially having on some customers. Please be assured we are doing everything within our power to resolve this incident as soon as possible.

We are continuing to work on a fix for this issue.

A fix has been implemented and we are monitoring the results.

This incident has been resolved.

At 10:03a.m. EST this morning we noticed that our visual search and input functionality was failing and traced the cause back to a 3rd-party vendor. We are waiting on their confirmation for a fix and we apologize for the inconvenience caused by this.

A fix has been implemented by our vendor and we are continuing to monitor it. All visual search and input operations should be working again as expected.

The fix from earlier today has been verified as stable. All systems are running normally now.

We have been experiencing issues with our Visual Search platform starting around 9:50am EST and are still investigating a potential fix. Any endpoints using inputs are also affected by this and our Explorer UI may experience slowness/loading issues as well. Apologies for any inconvenience caused by this.

We have identified the source of the issue and have implemented a fix for it. We will continue to monitor this to ensure that it is stable.

This incident has been resolved.

We have been experiencing issues with our Visual Search platform starting around 11:45am EST and are still investigating a potential fix. Apologies for any inconvenience caused by this.

All visual search operations are running as expected again as of 3:27pm EST. Postmortem to follow shortly.

Our engineering team has identified an issue related to Visual Search. During this time, users may receive error messages when attempting to issue Visual Search related API operations. We apologize for the inconvenience.

Additional Visual Search capacity has been added and our engineering team is monitoring request latency.

This incident has been resolved.

Around 6:00am EDT we started to experience some issues with our Custom Training, Visual Search and Explorer UI tools that would have likely interrupted access to them. This was resolved around 10:40am EDT and we are still investigating the cause/future solutions with one of our vendors. Apologies for any inconveniences that this caused.

Around 3:20p.m. EST we started to experience similar issues to yesterday's outage and are actively pursuing a fix. More information to come shortly.

A fix has been applied and all systems are running normally again.

This incident has been resolved.

Our Operations Team is currently looking into some issues with one of our third party vendors. Affected platforms: Custom Training, Visual Search, Explorer Tool We apologize for any inconvenience caused by this.

We have identified the root cause of today's issues and are currently working on a fix.

We have successfully restored service to the affected areas and are monitoring to ensure that it remains stable. Detailed post-mortem to follow.

This incident has been resolved.

Around 4:20pm EST we noticed a large spike in gateway timeout errors from our internal system logs. We are currently investigating this with a high priority and apologize for any inconvenience caused. Affected platforms: Visual Search, Custom Training, Input Uploads, Explorer Tool

Service was restored around 5:11 p.m. EST. We opened a ticket with our database vendor to determine the root cause and are still waiting to hear back from them.

We are still waiting on a root cause and will update with our postmortem accordingly. Apologies again for any inconveniences.

We are currently experiencing some issues with our Search platform and are looking into the root cause. Apologies for the inconvenience.

The issue has been resolved and all search operations should be working again. Post-mortem to follow.

We are currently investigating some database issues that occurred from a recent migration. As a result you may not be able to access your Clarifai applications in the interim.

The issue should now be resolved.

Around 6:25pm EST we discovered that database read/writes are randomly failing. We've opened a high priority ticket with our vendor and will continue to investigate.

We've identified that the problem is on the vendor's end and it appears that it has been resolved. We are continuing to monitor the situation.

The issue has now been resolved. We've confirmed that it randomly affected all database-related features (search, post inputs, etc.) between 5:28pm EST and 6:58pm EST.

Due to a required database update, we needed to pause all requests involving database changes from Nov 21 00:00:00 -- Nov 21 00:45:00 US Eastern Time. During the maintenance window, users were unable to add, modify or delete inputs and concepts, and any such calls would have resulted in an HTTP code 401, error code 39998, and an error message of “Input writes are disabled for maintenance. Please try again in a few hours.”. Predictions (public and custom) and searches with existing inputs remained unaffected during this period.

From 2:46pm EST to 4:22pm EST we were experiencing intermittent issues with our API, which resulted in some public model failures. A fix has been applied for this and we are continuing to investigate the root cause.

Around 4pm EST today we started seeing issues with our User Interface and subsequently, our API. We are currently looking into these and will update when we have more information. Apologies for the inconvenience caused by this.

We have identified the problem and have implemented a fix it. More information will be provided shortly on the cause.

The issue was resolved around 4:55pm EST

Around 12:55pm EST today, our public Logo Model stopped returning successful prediction results. We are still investigating the matter and should have more information shortly.

We instituted a fix for the issue at hand at 4:05pm EST and are monitoring its success. At the moment the model is making successful predictions again.

The code fix that we deployed has successfully fixed the issue and we have ensured that a similar case will be avoided in the future.