:wave: all. “new year” style question for those w...
# general
d
👋 all. “new year” style question for those with “mature” platforms: What fundamental feature do you wish had engineered into your platform earlier? Eg: granular cost attribution. Regular maintenance windows etc.
k
#1: Platform Team API (so everyone in the organization knows what we do & where to find us. Not a technical API, this: https://github.com/TeamTopologies/Team-API-template ) #2: Integrations with CMDB or other Enterprise Architecture repository. In long term, helps a lot. #3: Some "currency" exchange followed by fulfilment of Platform Service. It can be a "virtual" exchange of company-coins, then refleting in the budget - or real money transfered to Platform Team from Stream-aligned Team for a Platform Service #4: Automated success measure dashboards, like Provisioning Time etc.
d
Could you share some more thoughts around the “currency exchange”? Presumably there should also be some way to “spend” said currency - perhaps on headcount; tooling; training etc?
k
Sure - if you have a Platform Team, it should have some set of standardized services being provided to stream-aligned Team
Those services can have a price
Either in virtual coins, or - real transfers between Stream aligned team and Platform Team budget
Simply - Stream-aligned Team should somehow pay Platform Team
This drives some entrepreneurship into Platform Engineers to make services actually useful
s
It will really depend on organisation fit what the platform is missing, so for me it’s highly enforced standardisation, but my answer could be completely different for another company. Short lived teams, systems moving far beyond intended scope, regular organisational churn makes for an environment where strongly enforced standards to reduce impact of iterating on systems engineers are unfamiliar with would be the biggest impact IMO.
d
@Shane Dowling - any thoughts on the mechanisms for selecting and then enforcing said standards?
s
It’s a tricky non-standard problem(IMO), it really all depends on the company. To select and enforce standards in a mature platform, you need to involve cross-functional teams, including principal engineers, in the decision-making process. The aim being to ensure these standards align with both industry best practices and organisational needs. You’re also looking to get buy-in across the organisation. For enforcement, integrating automated compliance checks into CI/CD pipelines is a potential stick you can use(but needs strong backing). Depending on the organisation though, using automated reporting for real-time compliance visibility, coupled with regular training and documentation, ensuing everyone is on the same page might be enough depending on engineering maturity.
k
Simply speaking, the worst thing you can do is not to discover the cognitive load of the teams & their maturity :) I also regret not doing a proper Platform Observability (app observability was always a priority, which is fine - but some “business” metrics aggregation of how the Platform Services are working would have been useful sooner than I expected)