Cloud Cost Engineering and Open Source FinOps
Notes from the team building C3X. Cost estimation for Terraform, the economics of cloud infrastructure, and how to ship FinOps tooling without a SaaS gate.
GCP Region pricing differences: why location changes your bill
Google Cloud prices the same resources differently by Region, sometimes by 20 percent or more, and egress between Regions and to the internet varies too. Here is how GCP Region pricing works and how to pick Regions that balance cost, latency, and compliance.
GCP storage classes cost: Standard, Nearline, Coldline, Archive
Google Cloud Storage has four classes trading storage price against retrieval cost and minimum durations. Standard is cheapest to access, Archive cheapest to store. Here is how the classes are priced and how to pick the right one for each data set.
GitHub Actions minutes cost: what CI/CD really costs you
GitHub Actions bills by the minute for hosted runners, with multipliers for larger and non-Linux runners and per-GB charges for storage and data transfer. Here is how Actions minutes are priced, what the multipliers mean, and how to cut CI costs.
GKE Autopilot vs Standard cost: which Kubernetes mode is cheaper
GKE Standard bills for the nodes you provision; GKE Autopilot bills for the pod resources you request. Autopilot removes node management and idle waste but charges a premium per resource. Here is how the two pricing models compare and when each wins.
Glacier vs Deep Archive cost: which S3 cold tier to choose
S3 Glacier Flexible Retrieval, Glacier Instant Retrieval, and Glacier Deep Archive trade storage price against retrieval speed and cost. Deep Archive is the cheapest storage at around a dollar per TB per month. Here is how the cold tiers compare and when to use each.
Lambda provisioned concurrency cost: paying to kill cold starts
Provisioned concurrency keeps Lambda instances warm to eliminate cold starts, but you pay for that capacity around the clock whether it runs or not. Here is how the pricing works and when it is worth it.
Lambda vs Fargate cost: which serverless compute is cheaper
Lambda charges per request and per GB-second and scales to zero; Fargate charges per vCPU and GB by the second for continuously running tasks. The cheaper option depends entirely on your traffic pattern. Here is how to choose.
Lift and shift vs refactor cost: two paths to the cloud
Lift and shift moves workloads to the cloud as-is for speed, but often carries over-provisioned cost; refactoring for cloud-native services costs more up front but runs cheaper. Here is how to weigh the two.
LLM token cost explained: what you actually pay per call
Large language model APIs bill per token, split between cheaper input tokens and pricier output tokens, and the counts add up fast with long prompts and context. Here is how token pricing works and how to control it.
Log retention cost strategy: keeping logs without the surprise bill
Logs are cheap to write and expensive to keep forever. A retention strategy that tiers, expires, and archives logs by value cuts storage and ingestion cost without losing what you actually need. Here is how to build one.
Model hosting cost compared: API vs self-hosted inference
You can call a hosted model API and pay per token, or run your own model on GPU instances and pay for uptime. Each wins at a different scale. Here is how the two cost models compare and where they cross over.
NAT gateway surprise bill: the quiet networking cost
NAT gateways charge both an hourly fee and a per-GB data processing fee, so a chatty private subnet can rack up a large bill nobody expected. Here is why it happens and how to cut the cost.
Observability cost optimization: taming metrics, logs, and traces
Observability data (metrics, logs, and traces) is billed by volume, and unmanaged it can rival your compute bill. A cost-aware approach controls what you collect, how long you keep it, and how you sample. Here is the playbook.
On-prem to AWS migration cost: what the move actually costs
Migrating from a data center to AWS has a one-time project cost and a new ongoing run cost that replaces your capital spend. Understanding both, and how they compare to on-prem, is essential before you commit. Here is the breakdown.
OpenSearch cost optimization: taming a pricey cluster
Managed OpenSearch bills for data nodes, storage, and often dedicated master and warm nodes, so a large or over-provisioned cluster gets expensive fast. Here is how to right-size and tier it without losing search performance.
PrivateLink cost explained: private connectivity and its price
AWS PrivateLink gives private, secure access to services without traversing the public internet, but interface endpoints bill an hourly fee per endpoint per AZ plus a per-GB data processing charge. Here is how the cost works and when it pays off.
Prometheus vs managed monitoring cost: build or buy your metrics
Self-hosted Prometheus trades a per-metric SaaS bill for the cost of running and operating your own monitoring stack. Which is cheaper depends on scale, cardinality, and how much operational time you can spend. Here is the comparison.
RAG pipeline cost explained: where the money goes in retrieval
A retrieval-augmented generation pipeline spends on embeddings, a vector database, retrieval compute, and the LLM tokens for augmented prompts. Each stage has its own cost driver, and context size ties them together. Here is the full breakdown.
AWS X-Ray cost: per trace recorded and per trace scanned
X-Ray bills about $5 per million traces recorded and about $0.50 per million traces retrieved or scanned, with a free tier. Sampling is the main lever, since you rarely need every request traced. Here is how the tracing model prices.
Azure Traffic Manager cost: DNS queries plus health checks
Traffic Manager bills per million DNS queries plus a monthly fee per monitored endpoint, with external endpoints costing more to check. It is a cheap DNS-based global router. Here is how it prices and where cost hides with many endpoints.
GCP Secret Manager cost: per secret version and per access
Secret Manager bills about $0.06 per active secret version per location per month and about $0.03 per 10,000 access operations. It is cheap unless you store many versions or access secrets on every request. Here is the model and the caching lever.
AWS DataSync cost: about a cent and a quarter per gigabyte
DataSync bills about $0.0125 per GB of data copied, plus any underlying storage and transfer at the endpoints. Cost tracks how much you move, so incremental transfers and filters keep it down. Here is the simple model.
AWS Batch cost: free service, you pay for the compute it runs
AWS Batch adds no charge of its own; you pay for the EC2, Fargate, or Spot compute your jobs run on. Compute environment choice and Spot are the big levers. Here is how to run batch workloads cheaply.
AWS AppSync cost: per query, per real-time message, per connection
AppSync bills GraphQL query and data-modification operations per million, real-time subscription messages per million, and connection-minutes. Chatty clients and broad subscriptions drive the bill. Here is how the managed-GraphQL pricing works.