
Microcumulus
Unified OSS Dev Platform
Over the last 15 years, we've had some incredible experiences self hosting and growing open source tech stacks for myself and my employers. We'd love to popularize and centralize the things we've learned, while also getting to work more on something we're passionate about: freeing up people to think less about the infrastructure and more about the product, without paying much beyond the additional cost for compute, and built on open standards and tools so that you can still take most of your stack anywhere it makes sense. This includes being able to take it away from Microcumulus, though we hope to provide continued value to keep you using us.
The tech stack is generally as follows:
-
Gitlab CE as a cornerstone. GitLab offers 80% of the tools you need to build your software, track your features, and scale your teams. From issue boards to integrated CI/CD on k8s, it's got you covered and has tons of documentation on how to use it, plus a thriving community.
-
Kubernetes via kops: The best tool for deploying and managing clusters is still the kops tool. Our solution automates kops in a way that integrates with your own AWS accounts (so we don't foot the bill, and so that it's all still yours), while giving you the keys to walk away or manage things on your own. Currently we support AWS only, but KOPS supports a number of cloud providers, so new providers will require less effort while still providing most of the same options.
-
Backstage developer portal. Create your applications from some killer templates (zero to prod in minutes) with batteries included, and ready to be iterated on by your team (and/or ours) for standardization. Track software and teams (once you have teams) throughout the software lifecycle. Know where things are deployed and who currently owns them. Know which services use other services. Track your cloud costs by app, and have a single place for your team to search docs and APIs to ensure they get their data from the right place and don't need to duplicate effort.
-
Argo CD for GitOps (Optional?). Tracking the state of your current deployments can get complicated. GitLab offers some tooling for this, and that's where we started, but as you grow it is nice to have a separate registry, and Argo does a great job of this.
-
Keycloak for identity. The next most critical component of your business is keeping things secure. We deploy keycloak into your cluster and hook it up to every service we deploy, including the cluster itself for RBAC best practices. Every app template in backstage includes the option for keycloak authentication, which hooks everything up so you're ready to build your authenticated and secure service.
-
Vault and cert-manager for HTTPS. Stop worrying about certificates. We provide both a private and public CA, and the tools to easily get employees set up on the private CA, so that you can secure your application and so that you never have to worry about quotas causing outages, or transparency logs exposing private service names.
-
Open Telemetry (Prometheus, Grafana, Loki, Jaeger, AlertManager) for observability. Modern software stacks include the need to centralize observability data like logs, traces, and metrics. We set up a full blown stack with grafana dashboards built in, and have app templates all set up to report metrics, so you can focus on creating the right metrics and monitors to ensure availability. All of this is managed through an OTel collector so that if you decide you want to spend $$$$$ on another telemetry backend, you can.
-
Harbor (TBD) and Renovate bot (WIP). Track your dependencies, their vulnerabilities, and keep things up to date with open source security tooling that takes care of tracking and updating your dependencies on your behalf.
-
Rocketchat (TBD) for chat. Own your data and stop paying per-user costs every month. Build integrations on an open platform and ensure that your communications stay within your own walls.
-
Open source operators for data. Your app probably needs to track state, and we deploy some of the best open source operators for open source projects (like the Zalando postgres operator, others WIP) so that your data has somewhere to live, and gets backed up regularly with simple restoration. Baackstage templates come with options for persistence, so that you worry less about the best way to marry your code with the data for your language.
-
Strapi for quick iterative development and easy content management. We'll deploy strapi on your behalf and prep a dev environment so you can more easily manage your content types.
Why pay us?
You could do this all on your own. Why should you use us? We've taken the time to integrate it all out of the box, and get you up and running in tens of minutes.
We'll teach you how to run things. It's better for your customers and for us if you support your stack. Knowing the tech you depend on is a key enabler for us, and we want it to be for you too. Don't spend hours waiting for support to join your call; know your stack and know where to look from day 1 so that you can begin fixing things the second things go sideways.
We'll help you stay up to date, across the stack. Security updates happen, and we have experience rolling things out without impacting users. We'll teach you that too, if/when you want to take it over.
We also have built this on a budget conscious plan, using spot instances and minimum requirements where it makes sense, so that you can start cheap and grow with your user base.

About
We've spent years building platforms from the best OSS. We want more people to get to focus on their businesses and as little as possible on the infrastructure, but without spending $$$ on dozens of SaaS, and no lock-in.
Comment