# Scaleway: [Serverless Functions/Containers] - [nl-ams] - infrastructure upgrade (continuation)

- Provider: [Scaleway](https://outages.live/scaleway.md)
- Type: Maintenance
- State: scheduled
- Worst impact: Maintenance
- Started: 2026-11-25 05:00 UTC
- Components: nl-ams-1, nl-ams-2, nl-ams-3, Serverless Container, Serverless Functions, Elements - AZ - nl-ams-1, Elements - AZ - nl-ams-2, Elements - AZ - nl-ams-3, Elements - Products - Serverless Container, Elements - Products - Serverless Functions
- Official update: https://stspg.io/1qbj5hdgthy2

## Updates from Scaleway

### scheduled, 2026-11-19 09:51 UTC

On 11/25, from 7:00 AM UTC, we will continue the mandatory upgrade (started here: https://status.scaleway.com/incidents/f7lpf0j1c2sq) on our NL-AMS infrastructure. This upgrade will include changes allowing us to improve overall performance and prepare for the next features. While we tried our best to reduce impacts on users, some of them are inevitable. While we roll out the upgrade, expected impacts are, like previous maintenance: - Users' running functions/container instances will restart several times. As a result, requests could be buffered while new instances are started, causing "cold starts", even for functions/containers with min_scale set to 1. Functions/containers with no running instances won't be impacted. - 5xx errors while calling functions/containers. As we restart system components to perform the upgrade, some in-flight requests (especially long-running ones) might be terminated before completion. Users might also notice flakiness (slow requests, timeouts), but client

### resolved, 2026-11-19 09:48 UTC

No text.

### scheduled, 2026-11-19 09:42 UTC

On 11/25, from 7:00 AM UTC, we will continue the mandatory upgrade (started here: https://status.scaleway.com/incidents/f7lpf0j1c2sq) on our NL-AMS infrastructure. This upgrade will include changes allowing us to improve overall performance and prepare for the next features. While we tried our best to reduce impacts on users, some of them are inevitable. While we roll out the upgrade, expected impacts are, like previous maintenance: - Users' running functions/container instances will restart several times. As a result, requests could be buffered while new instances are started, causing "cold starts", even for functions/containers with min_scale set to 1. Functions/containers with no running instances won't be impacted. - 5xx errors while calling functions/containers. As we restart system components to perform the upgrade, some in-flight requests (especially long-running ones) might be terminated before completion. Users might also notice flakiness (slow requests, timeouts), but client
