Database clustering system and CNCF graduated project Vitess has recently become available in version 7.0, improving SQL support and stability amongst other things. The tool, which originated at YouTube, now sports VExec and Workflow commands to facilitate automation of resharding, materialization (rollups) and database split workflows.
It also comes with a framework built to allow dedicated connections, an addition that Vitess maintainer Deepthi Sigireddi said was “built with connection pooling in mind (in order to support high traffic applications), it was non-trivial to support dedicated connections. The framework that was built for this supports ‘reserve’, ‘taint’ and ‘release’ on connections. The use case that motivated this is that many popular frameworks change configuration settings per session and override the system settings.
Cloudflare eyes slice of AWS’ serverless pie
Web infrastructure provider Cloudflare last week introduced a private beta for Cloudflare Workers Unbound. The serverless platform aimed at software developers can be seen as a challenger to AWS Lambda, which Cloudflare flippantly dubbed as “legacy” in its announcement. To lure users of competing services, Cloudflare pledged “unparalleled flexibility, performance, security, ease of use, and pricing”.
Highlighted features include a decentralised approach, which is meant to reduce network latency, unthrottled CPU usage, zero nanosecond cold start times, and “no hidden fees”.
Control plane Kuma hits 0.7
Just a month after making its way into the CNCF sandbox, open source control plane / service mesh Kuma has upped its version count to 0.7. The latest iteration includes a new resource called zone which facilitates the addition of remote control planes, support for StatefulSets to run a service mesh across every service, as well as upgrades to the Proxy Template policy and kuma-dp.
New Relic restructures observability platform
Observability tooling provider New Relic has introduced a restructured version of its New Relic One platform in response to customer feedback. It is now made up of three products called Applied Intelligence, Full-Stack Observability, and a Telemetry Data Platform. While the latter takes care of collecting and analysing logs and metrics, Applied intelligence does anomaly detection, and Full-Stack Observability monitors infrastructure and apps.
To make the system easier to use and prevent users from having to switch between older and newer user interfaces, the components all come together under a new, modernised UI. According to the announcement, regular users shouldn’t have to do or learn anything new to get started with the new stuff. Devs who’ve toyed with the idea of giving New Relic One a try, meanwhile now get the chance to do so via a free tier, which allows 100GB of ingest per month and includes one Full-Stack Observability user license.
GitHub makes feature development more transparent
Repository management service GitHub is giving users some insight into what’s next for the platform, by providing a public roadmap. The new site is meant to help teams plan better and “share feedback earlier to influence what we’re building”. For Q3 2020 devs can, for example, look forward to policies for using external actions or code scanning for the cloud.
Oso goes into preview mode
Developers having trouble with integrating authorisation can now take open source policy engine oso for a spin. The open source project provides a declarative policy language for the authorisation logic, which is defined separately from the application, but executed inside of it, once the engine has been embedded. Oso is in developer preview right now, but already comes highly recommended from the likes of observability expert Charity Majors.