If your team can't answer "what depends on this?" in under 30 seconds, you're one bad delete away from an incident.
Graffin shows you the blast radius, the ownership, the drift, and the cost — in one graph.
https://t.co/5AxYoscio4
We synced Graffin to our own AWS environment.
2,767 resources. 2,134 with no management source. $993/month in untracked spend. 23% IaC coverage.
We built this to solve our own infra chaos. Here's what we found. 🧵#FinOps#DevOps
5 EKS clusters across our accounts.
Each one has its own dependency graph. Each one has resources nobody fully tracked. Each one is someone's job to know.
We have 2,134 resources with no management source. Most of them are in Kubernetes namespaces.
The problem is real.
What you don’t see is often what costs you the most.
A single forgotten instance quietly running.
$0.28/day → ~$100/year.
Multiply that across environments… it adds up fast.
If you’re not measuring it, you’re paying for it.
#cloud#aws#devops#finops
your cloud infrastructure has a story.
it's been saying it for months.
this one is about load balancers eating 40% of your spend.
kubernetes quietly creating 59 AWS resources.
4 orphans with no owner, no IaC, no reason to exist.
https://t.co/5AxYoscio4 reads it for you.
clicked on them. asked graffin what they were.
"these allow EC2 instances to communicate with
AWS Systems Manager without traversing the public internet"
no EC2 instances left.
no project left.
just three endpoints. billing forever.
your infrastructure has outgrown your memory
my AWS invoice said:
"Amazon Virtual Private Cloud $26.52"
that's it. no details. no names. no explanation.
i'd already deleted everything.
so what was still running?
ran a graffin sync.
instantly saw exactly where the cost was coming from.
3 vpc endpoints. still alive.
openclaw-vpce-ssm $0.24/day
openclaw-vpce-ssmmessages $0.24/day
openclaw-vpce-ec2messages $0.24/day
#openclaw
$21.60/month.