asencis Ltd

I'm looking for a robust tool that can analyse a statically generated Nuxt.js site on an SEO basis. Does anyone have any decent and reliable tools to understand SEO search ranking gaps? We are also looking at SEOptimer as another potential tool. We use Google Lighthouse to understand Speed Metrics, Performance, etc., but we would certainly like to understand the SEO within an automated build process.

So for our original marketing site, we recently moved from a pure Vue.js SPA (with some limited VueMeta and post-build performance optimization caked in). We were receiving Lighthouse performance reports in the 80s. Not bad. But with nuxt.js - we're at 99. Nuxt.js is the framework that should be replacing the boilerplate Vue.js SPA you've built your site with. Sure if you're building a Product Listing page in Wordpress, use Vue. But if you're building a whole marketing site - please don't use Vue! Our philosphy at asencis: build components with Vue - but scaffold them together with Nuxt into a coherant site.

We currently use Django Graphene v3 (link below) providing us with a GraphQL adhering to best principles. It is built upon the graphene python project and integrates seemlessly with Django. The project is well maintained, and is in active development. This library gives us excellent support for Relay based GraphQL, subscriptions, mutations and schema.

moon.js could be an interesting alternative for us to some marketing sites - we can build simply and quickly and ship the core at 7kb uncompressed! We are trialling switching to moon.js from Vue.js for our main website at https://asencis.com. So far we have shaved 45% off of the total bundle size.

We develop rapidly with docker-compose orchestrated services, however, for production - we utilise the very best ideas that Kubernetes has to offer: SCALE! We can scale when needed, setting a maximum and minimum level of nodes for each application layer - scaling only when the load balancer needs it. This allowed us to reduce our devops costs by 40% whilst also maintaining an SLA of 99.87%.

Again, for ease of use - we have a library that provides us with none of the often clumsy overheads of managing a CSS framework in-house. We can develop user-interfaces and user-experiences at rapid speed using a pattern library that we can trust. We also have implemented some of the features of the TailwindUI library. However, access to the TailwindUI components is a "paid-for-service", but certainly allows our junior developers to understand good, clean, "compentised" and accessible UI/UX.

prismic.io is the way forward for our development and CMS teams. We can integrate prismic.io in any of our applications via their very neat, self-documenting and well-organised GraphQL API. The WordPress RESTful API is a small miniture nightmare for developers, and sometimes overkill for a simple journal integration into an existing website.

What a debate to wade into - React vs. Vue.js. Prototyping of applications is much, much faster in Vue.js. React, I believe, has a much heavier developer learning experience - so hiring pure Javascript developers allows us to work in a much more framework agnostic way. However, React still has a place within our application stack - it's much more performant out of the box.

For bare-knuckle caching (optimising performance for user centered platforms) Memcache outperforms Redis by a fraction - a fraction that definitely counts at scale and volume. We use Redis for more stability when processing any volume of message queues. When compared to Redis, it uses less overhead memory.