Skip to main content

42 posts tagged with "Primary Site"

Release notes for self-hosted Foxglove Primary Site Helm chart.

View All Tags

Primary Site v0.0.77

This primary site release changes the name of the stream-server to query-server. It contains breaking changes, read the changelog to see if manual intervention is required for your deployment.

Changed
  • The stream-server container image has been renamed to query-server. New images will no longer be published under the stream-server name.
  • the streamService values namespace has been renamed to queryService. Values defined in the legacy streamService namespace will continue to work. streamService values override default values defined in the queryService namespace before templating.
  • breaking The stream-service application and deployment are removed, and replaced with an query-service deployment. Any auto-scaler which uses the stream-service deployment name as a scaleTargetRef will no longer have any effect. To upgrade safely:
    • Add a new autoscaler that targets the query-service deployment.
    • Upgrade and test your deployment.
    • Schedule a date in the future to remove the old autoscaler. After removing the old autoscaler, this chart should not be rolled back past this version.
  • The stream Kubernetes service is duplicated with the name query. New deployments should set up ingress to the query service, but the stream service remains for backwards compatibility.

Docs: About | Install | Upgrade

Primary Site v0.0.67

Added
  • inbox-listener: added MAX_FINALIZATION_WORKER_COUNT environment variable to configure the maximum worker count used to write files to the lake. This defaults to 40 and can be reduced to mitigate rate-limiting API errors from the lake bucket. When set to 0, the inbox listener uses one worker per topic in the input file.

Docs: About | Install | Upgrade

Primary Site v0.0.66

Fixes

  • Fixed an issue where the stream server would sometimes yield messages from earlier than expected when using the lastPerChannel replay policy.

  • Fixed an issue where too many concurrent connections to the underlying object store would result in resource exhaustion. There is now a default limit of 1024 concurrent connections.

Docs: About | Install | Upgrade

Primary Site v0.0.64

Fixes
  • Fixed an issue where input files with foxglove_device_name metadata would be interpreted as having no device.
Changed
  • Updated the default for SEQUENCE_NODE_MAX_BUFFERED_COUNT from 32 to 0.

Docs: About | Install | Upgrade

Primary Site v0.0.62

This release includes a fix for the stream service.

As with the last release, if you encounter issues with this image, you can opt out using the following configuration in your values.yaml file when upgrading the helm chart:

streamService:
  useLegacyImage: true

This option will be removed in future releases. Please report any issues encountered to [email protected].

Fixed
  • Fix AWS region selection so that the stream service can read from buckets in non-default regions.

Docs: About | Install | Upgrade