Not sure this is the right channel to ask, please ...
# general
s
Not sure this is the right channel to ask, please point me in the right direction if there's a better place where to talk about this. Our tech team is growing (50+ people) and we're starting to need a place where to centralize: • catalog repository • architectural documentation (how the services talk to each other, which service does what, etc...) • apis documentation (we have about 10 major services and internal apis we need to document, a mix of rest and graphql) • general best practices documentation (how to create e new service, development standards, etc...) • tools documentation (which 3rd party tools we're using for what) • centralising logging, monitoring, alerting (or at least dashboard that can help us to correlate infrastructure, application and business level logs, kpis and alerts) We're on GCP/Kubernetes, with application services in node & php We're partially using gcp logging/monitoring for infra & prometheus/grafana for application metrics and sentry for application errors. We're trying backstage right now, but it does not seem to be that easy to set up and use. We looked at mia-platform, but it does a little bit too much Anyone on a similar tech stack / situation that feels like sharing best practices and things you have done for building up your internal development platform? Or has anyone found something that is not too difficult to setup and customize that does not cost an arm and a leg?
a
what’s the goal? If you have so many services at 50 people you might want to reconsider overall architecture - i wouldn’t expect more than 10 total services at that size. So 5 documents that describe them
if you’re at 10-15 services an arch diagram that’s in figjam and just regularly reviewed…
have some high level principles on traffic flow, database migration principles etc
e
Hi For API documentation, my team build a web app to render swagger yaml file. This web app calls the swagger yaml file from gitlab using gitlab's API. For architectural documentation, my org is currently exploring on Anthos service mesh. As you already on GCP (thus, I assume you use GKE). Perhaps you would like to take a look into that
b
to document best practices, system architecture, purpose of 3rd party tools you are using, confluence or notion could be worth trying out.
s
Thanks @Ekky Kharismadhany
Sure, but I thought we were in a channel talking about best in class IDPs, and notion/confluence are may be decent for general documentation, but for sure there’s better for api docs and a few others of the points I mentioned…
For general documentation and api documentation I like https://readme.com For service discovery / dependency manager I like https://backstage.io For logging / monitoring / alerting I like https://newrelic.com But I can’t believe no one in this group has put something like this together in a consistent way… May be I need to build my own IDP? 😅
p
Hi @simone basso if you’re considering building your IDP, with backstage as a frontend and your services of choice made available to your developers for them to self-serve, I’d shamelessly like to suggest Kratix (kratix.io). It’s a flexible platform building framework so you can build the platform that fits your organisation, offering services on-demand and making it easier to manage day 2 and beyond. It’s definitely not too difficult to set-up and doesn’t cost an arm and a leg! One of our team did a demo at PlatformCon so you could take a look here:

https://youtu.be/x-YTUongRxk

e
Hi @simone basso, Sounds like you're facing some exciting challenges with your growing tech team! I'm Ella from Port, and I think we might have a solution that could help you centralize all your documentation and monitoring needs. Port offers a comprehensive internal developer portal that can centralize your: • Catalog repository • Architectural documentation • API documentation for both REST and GraphQL • General best practices documentation • Tools documentation • Centralized logging, monitoring, and alerting Given that you're on GCP/Kubernetes and using a mix of Node & PHP for your application services, Port can seamlessly integrate with your current setup. Plus, it's designed to be user-friendly and easy to set up, unlike some of the more complex solutions out there. Here are a few resources that might give you a better idea: • Using an internal developer portal for superior FinOps visibilityPort Adds Cloud Cost Management to its No-code Internal Developer Portal to Enable Simple FinOps for Platform EngineeringWhy AppSec Teams Need Internal Developer Portals Check the portal out yourself with our live, free, pre-populated live demo. If you'd like to learn more or see a demo, I'd be happy to chat further.
c
Hi @simone basso I'm Chris from CECG, we have a lot of expertise in what you look to be asking. We do IDP at quite a large scale which means we have our solutions for logging/metrics /capabilities etc. on very large multi-tenanted clusters. We are experts in platform engineering. Have a look at our site/Linkedin and if something catches your attention reach out to us. We have lots of interesting blogs on tech decisions as well on our site. https://www.cecg.io/
m
@simone basso Aside from the sales-based messages you’re getting on this thread, here’s my take (based on what I’ve built). There’s really no “one stop shop” right now that anyone has successfully built from an open-source perspective (it sounds like that’s what you’re looking for, but correct me if I’m wrong) other than Backstage. As Backstage stands right now, it’s all JS based. If you want to install it, JS. If you want to create a plugin, JS. If you want to import a plugin, JS. I “think” that’s what turns people off as they’re bound to doing it only one way. Other than that, I’d say Anthos since you’re already on GCP. In terms of Service documentation, oddly enough the thing that comes to mind is PlantUML. I know it’s old school, but it may be what you’re looking for.