Kong Outage History

Started Sep 29, 2026, 07:30 PM UTC· resolved Oct 01, 2026, 09:20 AM UTC

1. identified Sep 29, 2026, 07:30 PM UTC

   Customers may be unable to find newly-created items in Konnect using the Search function. Additionally, analytics and gateway manager filtering by newly-created objects may also be unavailable.

2. resolved Oct 01, 2026, 09:20 AM UTC

   This incident has been resolved.

Read the full incident report →

Started Aug 19, 2026, 08:39 PM UTC· resolved Aug 20, 2026, 01:16 AM UTC

1. identified Aug 19, 2026, 08:39 PM UTC

   Analytics queries may be unavailable or returning incomplete data in EU.

2. resolved Aug 20, 2026, 01:16 AM UTC

   This incident has been resolved.

Read the full incident report →

Started Aug 17, 2026, 09:33 PM UTC· resolved Aug 17, 2026, 09:51 PM UTC

1. resolved Aug 17, 2026, 10:10 PM UTC

   From 21:33z to 21:51z the Konnect Control Plane had an elevated error rate when updating some entities. This was due to a misconfiguration in our encryption policy evaluation that temporarily rejected all fields that required encryption.

Read the full incident report →

Started Aug 11, 2026, 02:31 PM UTC· resolved Aug 12, 2026, 06:00 AM UTC

1. investigating Aug 11, 2026, 02:31 PM UTC

   Some customers may notice some missing data from charts and graphs while this lag occurs. The team is working to fix the processing and back-fill any missing data.

2. resolved Aug 13, 2026, 04:34 PM UTC

   This incident has been resolved.

Read the full incident report →

Started Jul 24, 2026, 01:21 PM UTC· resolved Jul 24, 2026, 01:23 PM UTC

1. resolved Jul 24, 2026, 02:16 PM UTC

   We identified a brief period of elevated API errors in the EU region, from 13:21 - 13:23 UTC, resulting in HTTP 5xx responses across multiple Konnect services and potentially impacting actions taken on the Konnect UI. The underlying cause is still under investigation.

Read the full incident report →

Started Jul 13, 2026, 07:10 AM UTC· resolved Jul 13, 2026, 09:10 AM UTC

1. investigating Jul 13, 2026, 07:10 AM UTC

   Login to Konnect intermittently failing. Customers can retry login in order to gain access to the platform. API calls to Konnect are succeeding.

2. investigating Jul 13, 2026, 07:31 AM UTC

   We are continuing to investigate the issue.

3. investigating Jul 13, 2026, 08:44 AM UTC

   We are continuing to investigate the issue.

4. resolved Jul 13, 2026, 09:10 AM UTC

   This issue has been resolved.

Read the full incident report →

Started Jul 10, 2026, 01:55 AM UTC· resolved Jul 10, 2026, 02:41 AM UTC

1. investigating Jul 10, 2026, 01:55 AM UTC

   Customers are unable to visualize the platform usage analytics data. This does not impact data ingestion.

2. resolved Jul 10, 2026, 02:41 AM UTC

   This incident has been resolved.

Read the full incident report →

Started Jul 02, 2026, 01:56 PM UTC· resolved Jul 02, 2026, 02:01 PM UTC

1. monitoring Jul 02, 2026, 01:56 PM UTC

   We are monitoring an incident that resulted in an increased error rate in the Portal APIs. A fix has been applied, and error rates have decreased.

2. resolved Jul 02, 2026, 02:01 PM UTC

   This incident has been resolved.

Read the full incident report →

Started Jul 02, 2026, 08:30 AM UTC· resolved Jul 02, 2026, 09:08 AM UTC

1. resolved Jul 02, 2026, 09:51 AM UTC

   The Konnect control plane in the India was incorrectly rejecting all user tokens between 08:30am UTC to 9:08am UTC. A misconfiguration of the token validation URL was deployed, noticed by our monitoring, and rolled back. No dataplane traffic or proxy behavior would have been impacted. Impacts were: - API calls, including from tools like DecK and Terraform would receive 401 responses - Users on the Konnect UI may have been logged out if the UI made any control plane calls that resulted in a 401 response

Read the full incident report →

Started Jun 16, 2026, 06:43 PM UTC· resolved Jun 17, 2026, 01:21 PM UTC

1. investigating Jun 16, 2026, 06:43 PM UTC

   We are investigating reports of a small subset of users encountering application errors when accessing Konnect UI. Our team is actively investigating the issue and will provide updates as more information becomes available.

2. investigating Jun 16, 2026, 06:43 PM UTC

   We are continuing to investigate this issue.

3. monitoring Jun 16, 2026, 09:56 PM UTC

   A fix has been implemented and our team is actively monitoring the state.

4. resolved Jun 17, 2026, 01:21 PM UTC

   This incident has been resolved. The Konnect UI has recovered, and users should no longer experience application errors when accessing product areas in Konnect. The issue was only affecting a very small set of Konnect users. We will continue to monitor the service to ensure ongoing stability.

Read the full incident report →

Started Jun 15, 2026, 06:38 AM UTC· resolved Jun 15, 2026, 06:38 AM UTC

1. resolved Jun 15, 2026, 07:16 AM UTC

   Customers were unable to access the Konnect UI. Upon loading the interface, an error message reading "Could not fetch user session data - Please reload the page to try again" was displayed, preventing access to the platform. This issue affected all regions globally. Control plane configurations and data plane payloads using direct API access were not impacted during this time.

Read the full incident report →

Started Jun 09, 2026, 07:08 PM UTC· resolved Jun 09, 2026, 07:34 PM UTC

1. investigating Jun 09, 2026, 07:08 PM UTC

   Investigating incorrectly filtered results from the Developers list API

2. identified Jun 09, 2026, 07:15 PM UTC

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

3. monitoring Jun 09, 2026, 07:25 PM UTC

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

4. resolved Jun 09, 2026, 07:34 PM UTC

   This incident has been resolved.

Read the full incident report →

Started May 26, 2026, 03:41 PM UTC· resolved May 26, 2026, 05:12 PM UTC

1. monitoring May 26, 2026, 03:41 PM UTC

   From 14:00UTC to 15:38UTC some customers were unable to update custom plugins that required validation from the Konnect control plane. This would include schema and handler code validation. The konnect team has reloaded the validation service which allowed validations to continue again. The team is investigating the cause.

2. monitoring May 26, 2026, 03:41 PM UTC

   We are continuing to monitor for any further issues.

3. resolved May 26, 2026, 05:12 PM UTC

   The incident has been resolved. The cause is still being investigated and will be posted when available.

Read the full incident report →

Started May 20, 2026, 02:04 PM UTC· resolved May 20, 2026, 02:25 PM UTC

1. investigating May 20, 2026, 02:04 PM UTC

   Developers portals in the SG regions are not reachable. Existing app-registration and dataplane traffic continue to work. The team is investigating the cause.

2. monitoring May 20, 2026, 02:23 PM UTC

   The cause has been identified. The networking load balancer was unable to reach the gateway. Dev portals are now reachable.

3. monitoring May 20, 2026, 02:23 PM UTC

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

4. monitoring May 20, 2026, 02:24 PM UTC

   We are continuing to monitor for any further issues.

5. resolved May 20, 2026, 02:25 PM UTC

   This incident has been resolved.

Read the full incident report →

Started May 19, 2026, 09:22 PM UTC· resolved May 19, 2026, 09:55 PM UTC

1. investigating May 19, 2026, 09:22 PM UTC

   Konnect Control Plane is unavailable in the US East region. Customer dataplanes with valid configurations will remain up, but new dataplanes will not be able to pull configurations. The team is investigating.

2. monitoring May 19, 2026, 09:28 PM UTC

   The Konnect control plane in the US region has recovered. The team is continuing to monitor.

3. monitoring May 19, 2026, 09:28 PM UTC

   We are continuing to monitor for any further issues.

4. resolved May 19, 2026, 09:55 PM UTC

   The control plane in US continues to remain operational. The incident has been resolved.

Read the full incident report →

Started May 11, 2026, 03:06 AM UTC· resolved May 11, 2026, 03:19 AM UTC

1. investigating May 11, 2026, 03:06 AM UTC

   Analytics in the AU region is currently experiencing issues. No other regions are impacted at this time. Our engineering team is actively investigating the problem.

2. monitoring May 11, 2026, 03:09 AM UTC

   A fix has been deployed to the affected AU region. Analytics services are recovering, and we are actively monitoring the system to ensure stability.

3. monitoring May 11, 2026, 03:11 AM UTC

   We are seeing analytics services in AU region recovering.

4. resolved May 11, 2026, 03:19 AM UTC

   The issue affecting Analytics in the AU (ap-southeast-2) region has been resolved. The problem was caused by an incorrect node configuration, which has now been corrected and deployed. All services have recovered, and analytics data is returning as expected. No other regions were impacted.

Read the full incident report →

Started May 08, 2026, 08:00 PM UTC· resolved May 08, 2026, 08:00 PM UTC

1. resolved May 08, 2026, 08:00 PM UTC

   Kong proactively deployed an updated license to control planes ahead of an upcoming expiration to avoid any customer disruption. As a side effect of this update, data plane nodes were incorrectly displayed as out-of-sync in the Konnect UI. The data planes themselves continued operating normally with a valid configuration throughout. The team performed a progressive rollout that triggered a full configuration sync, which cleared the inaccurate status. All nodes returned to an in-sync state shortly after. No traffic was impacted at any point. Data planes continued routing requests normally and would receive configuration updates. The only customer-visible symptom was the inaccurate out-of-sync indicator in the UI. The full configuration sync step has been added to the standard runbook for license updates, so future license rotations will not produce this false positive. Engineering is also improving the out-of-sync detection logic to reduce false positives going forward.

Read the full incident report →

Started Apr 10, 2026, 06:07 PM UTC· resolved Apr 10, 2026, 09:11 PM UTC

1. investigating Apr 10, 2026, 06:07 PM UTC

   Some analytics data in the US may be experiencing data loss. We are currently investigating the cause of this issue and will provide updates as they become available.

2. identified Apr 10, 2026, 06:21 PM UTC

   We have identified the issue and are working on a fix

3. identified Apr 10, 2026, 06:41 PM UTC

   We have a fix and are rolling the change out. We will begin monitoring the situation closely and will provide another update once the rollout is complete.

4. monitoring Apr 10, 2026, 07:00 PM UTC

   We have rolled out the fix and will continue monitoring the situation. The system is now functioning as expected, but we are keeping a close eye on performance to ensure stability. We will provide another update if any issues arise.

5. resolved Apr 10, 2026, 09:11 PM UTC

   This incident has been resolved.

Read the full incident report →

Started Apr 09, 2026, 06:06 PM UTC· resolved Apr 09, 2026, 06:06 PM UTC

1. resolved Apr 09, 2026, 06:06 PM UTC

   Customers running 3.13 and below who use the OIDC plugin with `tls_client_auth_ssl_verify` unset would have seen this value change to `false` if they updated the config after the rollback of new defaults following the 3.14 release. The rollback incorrectly flipped the oidc plugin tls_client_auth_ssl_verify to false as a default, which was not one of the items recently switched to default true and has instead been defaulted true for some time. We have rolled out a fix to prod to change this default back to true. Updates to the OIDC plugin should once again keep this value set to true if not specifically defined.

Read the full incident report →

Started Apr 08, 2026, 04:20 PM UTC· resolved Apr 08, 2026, 09:33 PM UTC

1. investigating Apr 08, 2026, 04:20 PM UTC

   With the release of 3.14 and changes to default security settings for Kong’s secure by default initiatives, Konnect customers running dataplanes less than 3.14 and updating certain plugins without providing overrides to the new defaults began experiencing the following issues: Konnect would begin reporting that a default had been overridden that did not apply to the connected dataplane. This is a warning that Konnect gives when the configuration on Konnect control plane appears to have user-defined changes that do not apply to the dataplane version the customer is using. This message is provided to avoid a user configuring properties on a plugin that their dataplane would not utilize, to make it clear to users why a new field isn’t taking effect. Since our defaults changed, this caused the reporting in some cases to see this as an ‘override’ if the configuration didn’t match the new default, causing the message. This had no impact on dataplane configurations or behavior, but it was a confusing message, and we have removed it. The second and more impactful issue is the updating of default values in 3.14. After the 3.14 release, some fields like ssl_verify and hide_credentials in various entities started defaulting to true instead of false . This is causing customers who run a deck sync without these fields defined, will see their config values change from false to true which is an issue. Konnect is working on rolling back to the old default values. Once the default values are restored on the API, the next time the config is updated without the default values, the previous values will be applied. Plugins using ssl_verify: ace acme ai-aws-guardrail ai-azure-content-safety ai-llm-as-judge ai-proxy-advanced ai-rag-injector ai-rate-limiting-advanced ai-request-transformer ai-response-transformer ai-semantic-cache ai-semantic-prompt-guard ai-semantic-response-guard aws-lambda azure-functions basic-auth confluent confluent-consume datakit forward-proxy graphql-proxy-cache-advanced graphql-rate-limiting-advanced header-cert-auth http-log jwt-signer kafka-consume kafka-log kafka-upstream ldap-auth ldap-auth-advanced mtls-auth opa openid-connect proxy-cache-advanced rate-limiting rate-limiting-advanced request-callout response-ratelimiting saml service-protection tcp-log upstream-oauth Plugins using hide_credentials: Key-auth Key-auth-enc Basic-auth Hmac-authldap-auth Oauth2 Oauth2-introspection vault-auth (EE) ldap-auth-advanced (EE)

2. identified Apr 08, 2026, 04:20 PM UTC

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

3. resolved Apr 08, 2026, 09:33 PM UTC

   We have completed the rollback to the original default values. Customers applying their configurations without explicitly defining ssl_verify and hide_credentials will default to `false` again.

Read the full incident report →

Started Apr 02, 2026, 07:00 PM UTC· resolved Apr 02, 2026, 07:00 PM UTC

1. resolved Apr 02, 2026, 09:47 PM UTC

   On 4/2/2026. Konnect introduced a change as part of our 'secure by default' initiatives for the upcoming 3.14 release, which caused expected_hash syncing issues for some customers' data planes. For background, when a data plane connects to the Konnect control plane, the configuration is produced for that specific version of the Kong gateway. A configuration hash is calculated based off of the produced config, and stored as the expected config hash. When configuration changes occur, the control plane calculates the new configuration, broadcasts it to the connected data planes, and updates to the new expected hash. The issue here was the preparation of the control plane for supporting SHA256 to store credential hashes. When this change was introduced, it resulted in a minor structural difference in the produced configuration, but no actual change to the configurations occurred. This led to the same configuration yielding a different hash before and after the deploy. The way this could get a data plane to report 'out of sync' was the following: - A data plane is connected and in sync with DP hash and Expected Config Hash `a111111a` - The data plane is disconnected/reconnected, and since it is still in sync with DP hash and Expected Config Hash `a111111a`, a new config is not calculated - The change to embed SHA256 support in the control plane is deployed - A new data plane connects with no config loaded, and reports an empty config to the Control Plane - The control plane generates the slightly altered config, and ends up with a new config hash, which updates the expected config hash to `b22222b` - Since no configuration change occurred, the data planes already connected and on `1aaaaaa1` would not have been pushed a change event, and would still be running the same configuration as `b222222b` but the hash would not match, causing the DPs to appear out of sync Note: Once *any* change to the control plane configuration occurred, a new config event would have been delivered, and all previous data planes would have been updated to the hash based on the new structure. Additionally, if an impacted data plane were to disconnect and reconnect, they also would have been brought into sync with the new hash. The issue was the system failing to consider the new hash as an 'event' since it did not change configured proxy behavior. To address this in the short term, the Konnect engineering team has ensured structural changes like this which result in a config hash will trigger a configuration change event, even when no config has changed. To address this in the long term, Konnect is moving to a config version tracking methodology that will not be impacted by structural, non-behavioral changes as the hash is.

Read the full incident report →

Started Apr 01, 2026, 07:18 PM UTC· resolved Apr 01, 2026, 08:17 PM UTC

1. monitoring Apr 01, 2026, 07:18 PM UTC

   Creating, Reading, & Deleting PATs through the API & UI is degraded. This does not impact authentication with PATs.

2. monitoring Apr 01, 2026, 07:19 PM UTC

   The issue has been resolved and we are now monitoring.

3. resolved Apr 01, 2026, 08:17 PM UTC

   This incident has been resolved.

Read the full incident report →

Started Apr 01, 2026, 06:30 AM UTC· resolved Apr 01, 2026, 06:30 AM UTC

1. resolved Apr 02, 2026, 09:49 PM UTC

   On 4/2/2026. Konnect introduced a change as part of our 'secure by default' initiatives for the upcoming 3.14 release, which caused expected_hash syncing issues for some customers' data planes. For background, when a data plane connects to the Konnect control plane, the configuration is produced for that specific version of the Kong gateway. A configuration hash is calculated based off of the produced config, and stored as the expected config hash. When configuration changes occur, the control plane calculates the new configuration, broadcasts it to the connected data planes, and updates to the new expected hash. The issue here was the preparation of the control plane for supporting SHA256 to store credential hashes. When this change was introduced, it resulted in a minor structural difference in the produced configuration, but no actual change to the configurations occurred. This led to the same configuration yielding a different hash before and after the deploy. The way this could get a data plane to report 'out of sync' was the following: - A data plane is connected and in sync with DP hash and Expected Config Hash `a111111a` - The data plane is disconnected/reconnected, and since it is still in sync with DP hash and Expected Config Hash `a111111a`, a new config is not calculated - The change to embed SHA256 support in the control plane is deployed - A new data plane connects with no config loaded, and reports an empty config to the Control Plane - The control plane generates the slightly altered config, and ends up with a new config hash, which updates the expected config hash to `b22222b` - Since no configuration change occurred, the data planes already connected and on `1aaaaaa1` would not have been pushed a change event, and would still be running the same configuration as `b222222b` but the hash would not match, causing the DPs to appear out of sync Note: Once *any* change to the control plane configuration occurred, a new config event would have been delivered, and all previous data planes would have been updated to the hash based on the new structure. Additionally, if an impacted data plane were to disconnect and reconnect, they also would have been brought into sync with the new hash. The issue was the system failing to consider the new hash as an 'event' since it did not change configured proxy behavior. To address this in the short term, the Konnect engineering team has ensured structural changes like this which result in a config hash will trigger a configuration change event, even when no config has changed. To address this in the long term, Konnect is moving to a config version tracking methodology that will not be impacted by structural, non-behavioral changes as the hash is.

Read the full incident report →

Started Mar 19, 2026, 04:30 PM UTC· resolved Mar 19, 2026, 06:56 PM UTC

1. investigating Mar 19, 2026, 04:30 PM UTC

   We are investigating an issue with Dev Portals in ME region.

2. monitoring Mar 19, 2026, 04:36 PM UTC

   Dev Portals have recovered in the ME region. We are continuing to monitor the situation.

3. monitoring Mar 19, 2026, 04:39 PM UTC

   We are continuing to monitor for any further issues.

4. resolved Mar 19, 2026, 06:56 PM UTC

   This incident has been resolved.

Read the full incident report →

Started Mar 18, 2026, 08:13 PM UTC· resolved Mar 18, 2026, 08:45 PM UTC

1. investigating Mar 18, 2026, 08:13 PM UTC

   We are investigating possible degradation in the ME region that may be impacting Debugger and node listing.

2. identified Mar 18, 2026, 08:45 PM UTC

   We have identified the issue and are taking action to remove the affected instance.

3. monitoring Mar 18, 2026, 08:46 PM UTC

   We have corrected the issue and are actively monitoring the situation.

4. resolved Mar 19, 2026, 07:38 PM UTC

   This incident has been resolved.

Read the full incident report →