Live outage feed

Outages and recoveries from the services you depend on.

Earlier incidents

40 incidents

Back to latest

Feb 3, 2026

  1. ClickHouse CloudElevated error rate in AWS

    We have confirmed that all systems are stable and operating normally. The networking configuration has been fully restored across all regions, and we have implemented additional monitoring and safeguards to prevent recurrence. We will continue to monitor systems closely over the coming days as part of our standard post-incident procedures, and share an RCA as soon as possible.

    Resolved

Feb 2, 2026

  1. LeverAutomation Workflow Trigger Delays

    We want to share an update on a recent automation issue, including what happened, how it was resolved, and the steps we’ve taken to prevent it from happening again. ## Customer Impact Between **January 21 and January 23, 2026**, some customers experienced delays in automated workflow execution. Automation workflows entered a _“waiting”_ state and were processed behind schedule rather than in real time. While no data was lost, workflow execution was slower than expected, which may have delayed downstream actions that depend on automation completion. Service levels were fully restored by **January 23, 2026 at 08:54**, and workflows have been operating in real time since that point. ## Root Cause We had a single instance that had accumulated a very large number of automation workflows configured to trigger on the same event type. Each time an application event occurred, all of these workflows were evaluated simultaneously. This configuration resulted in **tens of thousands of unnecessary database queries per event**, which significantly increased system load. As volume grew, the automation system was no longer able to process workflows at the rate they were being generated, causing a backlog and delays for all automation tasks. In short: * The automation system encountered an unusually high and inefficient workload pattern. * Existing safeguards and scaling assumptions were insufficient for this edge case. ## Resolution Once the issue was identified, we implemented multiple performance optimizations to reduce unnecessary processing and improve throughput, including: * Skipping workflows that did not apply to the triggering event before running expensive queries. * Reducing redundant queries where only a single result was required. * Improving the overall efficiency of the automation workflow. These changes significantly reduced system load, allowed processing queues to catch up, and restored real‑time workflow execution. Service levels returned to normal by **January 23, 2026**, and the system has remained stable since. ## Preventative Actions To prevent similar incidents in the future, we have taken and are continuing the following actions: * **Improved monitoring and alerting** for automation workflow latency, enabling earlier detection before customer impact. * **Hardened automation workflow performance**, ensuring the system can safely handle large numbers of workflows without unnecessary processing. * **Reviewed and updated testing strategies** to better reflect real‑world production workloads and edge cases. * **Ongoing optimization work** to further improve the scalability and resilience of the automation system. * **Evaluation of capacity and scaling behavior** to ensure consistent performance during peak and non‑peak hours. These measures are already in place or actively underway and are designed to ensure automation workflows continue to operate reliably as usage grows. We appreciate your trust and patience, and we’re committed to continuously improving the reliability and performance of the automation experience.

    Resolved

  2. BufferTrouble loading Buffer

    Due to internal work, Buffer encountered issues that caused some users to experience trouble loading the dashboard. An errant change was reverted, and functionality was restored. The trouble lasted for approximately 10 minutes.

    Resolved

  3. FigmaFigma Make Publishing Disruption

    From 9:58 PM UTC on Jan 27 to 2:00 AM UTC on Feb 1, some users experienced issues when prompting in Figma Make or Sites with code components. This issue has been resolved.

    Resolved

Feb 1, 2026

  1. BufferIssue Publishing Image Posts to Threads

    The issue publishing image posts to Threads has been fully resolved, and all systems are now operating normally. We appreciate your understanding and apologies for any disruption this may have caused.

    Resolved

  2. PlivoVoice and Zentrunk partial outage

    We observed elevated latency for approximately 10 minutes, which may have briefly affected performance for some users. The underlying issue was identified and addressed, and the system has since returned to normal operation.

    Resolved

Older incidents