Platform review
Find out what to fix first
An outside look at how your platform is built and run. In two weeks we go through your Kubernetes, Terraform and pipelines, talk to the engineers who use them, and give you a ranked list of what to fix and in what order.
- Duration
- Two weeks
- Price
- Fixed, agreed before we start
- Your team's time
- About half a day in total, spread across a few engineers
- Access
- Read-only. We never need write access to production for a review.
What we look at
The whole path from commit to production
We agree the scope with you first, but most reviews cover these six areas. We look at the code and at how things actually run, because the two are rarely the same.
- Infrastructure as code
- Drift between Terraform and reality, state handling, module structure and how changes are reviewed and applied.
- Kubernetes clusters
- Versions and upgrade path, node groups, resource requests, add-ons and how consistently clusters are built.
- Delivery pipelines
- How code gets from commit to production, how long it takes, how deploys roll back and what CI is allowed to touch.
- Security and access
- Long-lived keys, secrets handling, who is cluster-admin, network exposure and the IAM behind it all.
- Cost
- Over-provisioned workloads, idle capacity, extended support fees and the network charges nobody expected.
- Reliability and recovery
- Alerting that people act on, backups that have been restored, and how long a rebuild from code would take.
How it runs
Light on your team's time
Before we start
Scope and access
A short call to agree what's in scope, then read-only access to your repositories, cloud accounts and clusters.
Week 1
Look and listen
We read the code and config, look at how things actually run, and spend an hour or so with a few of your engineers.
Week 2
Write it up
We rank every finding by impact and effort, and check the quick wins with your team so there are no surprises.
End of week 2
Walk through it
A session with your team to go through the report, answer questions and agree what to tackle first.
What you get
A plan, not a slide deck
- A written report with every finding ranked by impact and effort
- A short list of quick wins your team can do in the next few weeks
- A suggested order for the bigger changes, roughly a quarter's worth
- A walkthrough session with your engineers and anyone else who needs to hear it
- A follow-up call a month later to answer questions as you make changes
The report is yours to keep and act on however you like. Plenty of teams take it and do the work themselves. If you'd like help with some of it, we can talk about that afterwards, but there's no obligation.
A good fit if
- You run Kubernetes on AWS, Google Cloud or Azure, or are about to
- The platform works, but it's slower, more expensive or more fragile than it should be
- You want an outside view before a big decision, a funding round, an audit or a new hire
Need something bigger, like a full platform build or a migration? Have a look at ourother services.
Book a platform review
Tell us roughly what you run and what's worrying you. We'll reply personally with dates and a fixed price.