Section Release Notes
Platform improvements and other notable updates
White Labeled CNAME Domains, UX Improvements, Better Log Sampling
October 25, 2021
- White labeled CNAME domains now available
- Improved UX when working in Advanced Config
- Updates to Egress config requirements
- Introduced new log sampling algorithm to deliver more useful results
Below are some highlights of recent releases.
White labeled CNAME domains now available
Section customers now have the option to mask the default Section CNAME DNS server name (i.e. c.section.io) and replace it with their own specified domain name (e.g. c.example.com). Currently, all customers are provided with a CNAME record for each domain they provision on Section in the
c.section.io zone. With this new white label feature, customers can use their own domains to provide better brand continuity to their customers while masking the underlying Section infrastructure. If no custom domain is provided, then the Section CNAME record will be provisioned by default.
Improved UX when working in Advanced Config
We made improvements to the Git branch logic in the Section Console to better track and display the active working environment (e.g. Staging, Production). This update helps ensure that code commits are pushed to the intended branch when working within the Advanced Config settings.
Updates to Egress config requirements
We’ve updated the handling of Egress pods to remediate potential incidents associated with placeholder configurations being used when responding successfully on readiness probes. With these updates, Egress pods in production PoPs now only advertise as traffic-ready once they’ve parsed and applied egress configurations.
Introduced new log sampling algorithm to deliver more useful results
Section uses log sampling for high volume environments to help offset the impact of resource-intensive processes (transmission, storage, search, etc.). For Modules with the same sampling rate configured, the new algorithm results in all logs for a given request all being accepted together, or all being rejected together, because their
section_io_id is equal. This will allow a log viewer to see the full journey of a request through the platform even when aggressively sampled.
Previously, the algorithm was random, so a log viewer would typically only see subsets of the logs for a given request which could make diagnostics tricky. Logs without
section_io_id values (e.g. some Module error logs) will still revert to random sampling.