Langfuse

Failure mode: Exhausted memory, disk, CPU, queues, connection pools or quotas. Root cause: The vendor does not say why it happened. Tagged by claude-haiku-5-5 with low confidence (taxonomy v0.2).

1. investigating 2026-10-21 07:46 UTC

   Slow traces data serving across API and UI

2. identified 2026-10-21 08:31 UTC

   Slow clickhouse read replicas affecting performance across the service. Draining and replacing them.

3. resolved 2026-10-21 08:45 UTC

   All replicas look healthy again

4. monitoring 2026-10-21 08:56 UTC

   Looks like some timeout are still occurring. Monitoring.

5. resolved 2026-10-21 09:39 UTC

   should be finally resolved

Excerpts are shortened. Read the full incident on Langfuse's status page.

Failure mode: A code release introduced the failure. Root cause: The vendor does not say why it happened. Tagged by claude-haiku-5-5 with low confidence (taxonomy v0.2).

1. investigating 2026-10-20 17:03 UTC

   We are seeing increased error rates for requests in Halo (Langfuse assistant)

2. identified 2026-10-20 17:11 UTC

   We identified the issue and a fix is currently being deployed.

3. resolved 2026-10-20 17:50 UTC

   Issue with Langfuse Halo (assistant) is fixed.

Excerpts are shortened. Read the full incident on Langfuse's status page.

Failure mode: Data delayed, stale, lost or corrupted, or needing replay after recovery. Root cause: The vendor does not say why it happened. Tagged by claude-haiku-5-5 with medium confidence (taxonomy v0.2).

1. identified 2026-10-20 15:54 UTC

   Degraded data serving in EU due to unhealthy clickhouse replica. Draining traffic.

2. monitoring 2026-10-20 16:01 UTC

   We should have recovered

3. resolved 2026-10-20 16:20 UTC

   Have been stable

Excerpts are shortened. Read the full incident on Langfuse's status page.

Failure mode: Exhausted memory, disk, CPU, queues, connection pools or quotas. Root cause: The vendor does not say why it happened. Tagged by claude-haiku-5-5 with low confidence (taxonomy v0.2).

1. investigating 2026-10-19 10:01 UTC

   Degraded public API performance in US. Investigating.

2. resolved 2026-10-19 10:46 UTC

   Problematic workload was identified and isolated.

Excerpts are shortened. Read the full incident on Langfuse's status page.

Failure mode: Exhausted memory, disk, CPU, queues, connection pools or quotas. Root cause: Not enough capacity for the load. Tagged by claude-haiku-5-5 with medium confidence (taxonomy v0.2).

1. identified 2026-09-30 09:10 UTC

   S3 backpressure leads to S3 "please slow down" errors across projects with cm* prefix

2. monitoring 2026-09-30 09:43 UTC

   Offending workload was isolated, performance should be back to normal for other projects

3. resolved 2026-09-30 15:30 UTC

   All projects ingesting as normal

Excerpts are shortened. Read the full incident on Langfuse's status page.

Failure mode: The text does not say enough to tell. Root cause: The vendor does not say why it happened. Tagged by claude-haiku-5-5 with low confidence (taxonomy v0.2).

1. investigating 2026-09-23 13:36 UTC

   We're observing elevated error rates across our In-App agent across multiple regions. Our team is investigating.

2. resolved 2026-09-23 13:43 UTC

   The issue has been resolved.

Excerpts are shortened. Read the full incident on Langfuse's status page.

Failure mode: Data delayed, stale, lost or corrupted, or needing replay after recovery. Root cause: The vendor does not say why it happened. Tagged by claude-haiku-5-5 with medium confidence (taxonomy v0.2).

1. investigating 2026-09-16 20:20 UTC

   We're investigating degraded performance in the ingestion flow

2. resolved 2026-09-16 21:31 UTC

   Issue is resolved and OTEL ingestion pipeline is back to regular ingestion speed.

Excerpts are shortened. Read the full incident on Langfuse's status page.

Failure mode: Data delayed, stale, lost or corrupted, or needing replay after recovery. Root cause: The vendor does not say why it happened. Tagged by claude-haiku-5-5 with medium confidence (taxonomy v0.2).

1. investigating 2026-09-14 02:51 UTC

   We currently experience an ingestion delay of old SDKs by 3 hours. This is due to performance issues in our batch job which moves data into the v4 data model ( ). We currently work on improving the batch job. All v4 compatible SDKs are not impacted by this. Upgrading SDKs is...

2. investigating 2026-09-14 03:52 UTC

   We considerably increased throughput of the batch job to write data into v4 data model. We are currently catching up processing data.

3. monitoring 2026-09-14 05:14 UTC

   We're making consistent progress in our backlog and are currently seeing delays of about 60 minutes.

4. resolved 2026-09-14 06:10 UTC

   The incident has been resolved.

Excerpts are shortened. Read the full incident on Langfuse's status page.

Failure mode: Data delayed, stale, lost or corrupted, or needing replay after recovery. Root cause: The vendor does not say why it happened. Tagged by claude-haiku-5-5 with medium confidence (taxonomy v0.2).

1. investigating 2026-09-08 17:26 UTC

   We're investigating degraded performance in the ingestion flow

2. resolved 2026-09-08 19:00 UTC

   OTEL ingestion latency spike is resolved

Excerpts are shortened. Read the full incident on Langfuse's status page.

Failure mode: Traffic surge or denial-of-service attack. Root cause: The vendor does not say why it happened. Tagged by claude-haiku-5-5 with medium confidence (taxonomy v0.2).

1. investigating 2026-09-02 16:02 UTC

   We are observing OTEL ingestion backlog in US region

2. identified 2026-09-02 16:24 UTC

   A deployment rollout coincided with a traffic spike, resulting in an unusually large backlog. Adjusting our autoscaling to absorb the backlog faster.

3. monitoring 2026-09-02 16:49 UTC

   Backlog is almost completely gone. Ingestion delay is < 1 min.

4. resolved 2026-09-02 16:56 UTC

   Completely back to normal

Excerpts are shortened. Read the full incident on Langfuse's status page.

Failure mode: Exhausted memory, disk, CPU, queues, connection pools or quotas. Root cause: The vendor does not say why it happened. Tagged by claude-haiku-5-5 with low confidence (taxonomy v0.2).

1. monitoring 2026-08-25 15:23 UTC

   Ingestion delay of up to multiple minutes on the ingestion path. We have scaled up and are investigating the root cause.

2. resolved 2026-08-25 16:28 UTC

   Issue resolved. Latency is back to normal levels.

Excerpts are shortened. Read the full incident on Langfuse's status page.

Failure mode: Exhausted memory, disk, CPU, queues, connection pools or quotas. Root cause: The vendor does not say why it happened. Tagged by claude-haiku-5-5 with medium confidence (taxonomy v0.2).

1. monitoring 2026-08-25 10:16 UTC

   The legacy ingestion queue is experiencing latency degradation up to 5 minutes. We have scaled up our infrastructure in response and are monitoring the situation. We recommend to switch to the new OTEL ingestion queue.

2. resolved 2026-08-25 13:42 UTC

   Issue resolved since 12:30. Legacy ingestion pipeline is fully operational.

Excerpts are shortened. Read the full incident on Langfuse's status page.

Failure mode: Exhausted memory, disk, CPU, queues, connection pools or quotas. Root cause: The vendor does not say why it happened. Tagged by claude-haiku-5-5 with medium confidence (taxonomy v0.2).

1. investigating 2026-08-20 08:06 UTC

   Evaluations are taking longer to execute in EU

2. monitoring 2026-08-20 09:40 UTC

   After concurrency adjustments execution delay is back to normal. Monitoring.

3. resolved 2026-08-20 09:55 UTC

   All is back to normal

Excerpts are shortened. Read the full incident on Langfuse's status page.

Failure mode: Exhausted memory, disk, CPU, queues, connection pools or quotas. Root cause: Not enough capacity for the load. Tagged by claude-haiku-5-5 with medium confidence (taxonomy v0.2).

1. monitoring 2026-08-19 10:30 UTC

   We have scaled our database infrastructure to match the throughput demand

2. monitoring 2026-08-19 13:19 UTC

   Latest major SDK versions are not impacted. Find upgrade docs [here](https://langfuse.com/docs/observability/sdk/upgrade-path).

3. resolved 2026-08-19 15:10 UTC

   Back to normal propagation delay

Excerpts are shortened. Read the full incident on Langfuse's status page.

Failure mode: Exhausted memory, disk, CPU, queues, connection pools or quotas. Root cause: Not enough capacity for the load. Tagged by claude-haiku-5-5 with medium confidence (taxonomy v0.2).

1. investigating 2026-08-18 12:06 UTC

   Eval execution in EU region is taking longer than usual

2. identified 2026-08-18 12:27 UTC

   evaluation throughput was overwhelmed by a slow tenant/model workload, with lock-expiry retries multiplying work; the concurrent worker rollout prolonged the incident.

3. monitoring 2026-08-18 12:47 UTC

   our autoscaling responded slower than expected, but responded nonetheless. the backlog is draining

4. resolved 2026-08-18 13:25 UTC

   Steady throughput for the last 30 minutes. Delay < 5min.

Excerpts are shortened. Read the full incident on Langfuse's status page.

Failure mode: Data delayed, stale, lost or corrupted, or needing replay after recovery. Root cause: The vendor does not say why it happened. Tagged by claude-haiku-5-5 with medium confidence (taxonomy v0.2).

1. investigating 2026-08-15 17:28 UTC

   Currently the ingestion for the events table is delayed in the US environment. This happens only for data from non-latest major SDK versions. We are investigating this and have started a mitigation. The data should be up to date soon.

2. resolved 2026-08-15 18:05 UTC

   We incurred a temporary delay on the legacy ingestion path for v4. This delay has been resolved. OTEL ingestion has not been impacted.

Excerpts are shortened. Read the full incident on Langfuse's status page.

Failure mode: Exhausted memory, disk, CPU, queues, connection pools or quotas. Root cause: The vendor does not say why it happened. Tagged by claude-haiku-5-5 with medium confidence (taxonomy v0.2).

1. monitoring 2026-08-12 11:56 UTC

   We are seeing an increase in the eval execution. We have scaled up our infrastructure and are monitoring.

2. monitoring 2026-08-12 12:39 UTC

   Previous latency increase of the eval execution is resolved. Slight delays may still occur.

3. resolved 2026-08-12 13:37 UTC

   The issue is resolved now

Excerpts are shortened. Read the full incident on Langfuse's status page.

Failure mode: Exhausted memory, disk, CPU, queues, connection pools or quotas. Root cause: The vendor does not say why it happened. Tagged by claude-haiku-5-5 with low confidence (taxonomy v0.2).

1. investigating 2026-08-03 18:25 UTC

   We're currently observing about 10min ingestion delays. Our team is investigating.

2. investigating 2026-08-03 18:56 UTC

   We see an improvement and the current delay is below 5min. We continue to investigate and scale.

3. resolved 2026-08-03 19:03 UTC

   This issue is resolved and we're back to regular processing times.

Excerpts are shortened. Read the full incident on Langfuse's status page.

Failure mode: The text does not say enough to tell. Root cause: The vendor does not say why it happened. Tagged by claude-haiku-5-5 with low confidence (taxonomy v0.2).

1. investigating 2026-07-14 19:12 UTC

   We experience occasional API errors with 502 Bad Gateway. The team is investigating the root cause currently.

2. resolved 2026-07-14 20:19 UTC

   We resolved the performance bottleneck. All APIs are responding as expected as of now.

Excerpts are shortened. Read the full incident on Langfuse's status page.

Failure mode: The text does not say enough to tell. Root cause: The vendor does not say why it happened. Tagged by claude-haiku-5-5 with low confidence (taxonomy v0.2).

1. investigating 2026-07-02 07:43 UTC

   We're observing elevated latencies and error rates across read APIs and some frontend routes. Ingestion is not affected.

2. resolved 2026-07-02 08:03 UTC

   The issue has been resolved.

Excerpts are shortened. Read the full incident on Langfuse's status page.

Failure mode: Data delayed, stale, lost or corrupted, or needing replay after recovery. Root cause: The vendor does not say why it happened. Tagged by claude-haiku-5-5 with medium confidence (taxonomy v0.2).

1. investigating 2026-06-26 09:32 UTC

   We're currently experiencing elevated ingestion times in the EU environment. Some events are processed with a delay of about 15 minutes. Our team is investigating.

2. monitoring 2026-06-26 09:43 UTC

   We found a patch and currently work through the backlog. Things should recover within the next 10min.

3. resolved 2026-06-26 09:51 UTC

   The incident has been resolved.

Excerpts are shortened. Read the full incident on Langfuse's status page.

Failure mode: Orchestration, control plane, service discovery or load balancing failed. Root cause: A human mistake during operations. Tagged by claude-haiku-5-5 with medium confidence (taxonomy v0.2).

1. monitoring 2026-06-16 17:36 UTC

   A hiccup during a planned database maintenance. Identified and resolved. Presently monitoring.

2. resolved 2026-06-16 18:12 UTC

   All latencies are back to normal

Excerpts are shortened. Read the full incident on Langfuse's status page.

Failure mode: Exhausted memory, disk, CPU, queues, connection pools or quotas. Root cause: Not enough capacity for the load. Tagged by claude-haiku-5-5 with medium confidence (taxonomy v0.2).

1. investigating 2026-06-10 14:47 UTC

   We currently have a bottleneck in our ingestion pipeline which causes a delay of ingested data of 12 minutes. We currently work to scale the pipeline.

2. resolved 2026-06-10 15:16 UTC

   We scaled our ingestion and are processing data in time.

Excerpts are shortened. Read the full incident on Langfuse's status page.

Failure mode: A schema or data migration caused the failure. Root cause: A human mistake during operations. Tagged by claude-haiku-5-5 with low confidence (taxonomy v0.2).

1. investigating 2026-06-08 20:55 UTC

   We just executed a database upgrade which caused a downtime of 5 minutes.

2. resolved 2026-06-08 20:55 UTC

   Upgrade successfully resolved.

Excerpts are shortened. Read the full incident on Langfuse's status page.

Failure mode: Exhausted memory, disk, CPU, queues, connection pools or quotas. Root cause: Not enough capacity for the load. Tagged by claude-haiku-5-5 with medium confidence (taxonomy v0.2).

1. investigating 2026-06-08 18:57 UTC

   We observe that our api/public/media endpoint currently returns 500 status errors as our underlying database has performance issues.

2. investigating 2026-06-08 19:54 UTC

   We scaled our underlying infrastructure and are processing media uploads without errors.

3. resolved 2026-06-08 19:55 UTC

   We scaled our infrastructure and are processing media uploads without errors.

Excerpts are shortened. Read the full incident on Langfuse's status page.