Resolved incident

Keeper API website load time and API errors in US region

Keeper · Critical · Started

Duration
1h 1m
Severity
Critical
Status
Postmortem

Summary

Critical · Postmortem

On Tues, Sept 24 at 6:06AM PDT, the DevOps Engineering team started receiving notifications regarding API slowness and HTTP 504 errors coming from the Keeper AWS infrastructure. After an investigation, it was determined that our NGINX instances were throwing the errors due to an overload of network traffic. It was then determined that the network traffic increase was caused by an update to the Keeper Desktop application which was published on Monday, the day before. The DevOps team scaled the AWS infrastructure several times over the period of 60 minutes to address the issue. The issue was fully stabilized and resolved by 7:20AM PDT. After a full analysis of the product update process, the engineering team has decided to change the update mechanism of the Keeper Desktop application to utilize a new distributed content delivery network which does not impact the other Keeper AWS infrastructure. This change will be implemented over the next 2 weeks and in preparation of the next Keeper Desktop application update.

View official incident ↗

Event timeline

How this incident unfolded, from the official status page.

  1. Investigating

    Keeper

    We are currently investigating reports of slow server response times and request timeouts

  2. Identified

    Keeper

    Root cause has been identified and we are actively working on a resolution.

  3. Monitoring

    Keeper

    A resolution has been rolled out - the team is monitoring system performance.

  4. Monitoring

    Keeper

    We are continuing to monitor for any further issues.

  5. Resolved

    Keeper

    This incident has been resolved.

  6. Postmortem

    Keeper

    On Tues, Sept 24 at 6:06AM PDT, the DevOps Engineering team started receiving notifications regarding API slowness and HTTP 504 errors coming from the Keeper AWS infrastructure. After an investigation, it was determined that our NGINX instances were throwing the errors due to an overload of network traffic. It was then determined that the network traffic increase was caused by an update to the Keeper Desktop application which was published on Monday, the day before. The DevOps team scaled the AWS infrastructure several times over the period of 60 minutes to address the issue. The issue was fully stabilized and resolved by 7:20AM PDT. After a full analysis of the product update process, the engineering team has decided to change the update mechanism of the Keeper Desktop application to utilize a new distributed content delivery network which does not impact the other Keeper AWS infrastructure. This change will be implemented over the next 2 weeks and in preparation of the next Keeper Desktop application update.