756 articles on cost optimization — what drives the cost, how it is priced, and where the savings actually are.
Stopping an EC2 instance halts compute charges but keeps EBS billing; hibernating preserves memory to disk for fast resume but stores that memory on EBS, adding cost. For part-time workloads, knowing which to use and what still bills is worth real money. Here is the comparison.
Development, test, and staging environments rarely need to run outside business hours, yet many run 24/7. Scheduling them off nights and weekends cuts their cost by roughly 65 to 75 percent with no loss of function. Here is how to do it and what it saves.
Lambda cost per million invocations depends on two levers, the flat per-request fee and the per-GB-second duration charge, and the duration part usually dominates. Working through real numbers shows exactly where the cost goes and which lever to pull. Here is the breakdown.
Two numbers decide whether Savings Plans save money or waste it: coverage (how much of your usage the commitment covers) and utilization (how much of the commitment you actually use). Over-commit and utilization drops; under-commit and coverage is low. Here is how to balance them.
EC2 offers four ways to pay, on-demand, Spot, Savings Plans, and Reserved Instances, each with a different discount, commitment, and risk profile. Using the right one for each part of a workload is worth 50 to 80 percent off the naive on-demand bill. Here is the full comparison.
App Runner and Fargate both run containers without managing servers, but they price and scale differently. App Runner bundles more of the operational layer at a per-resource rate, while Fargate gives finer control and can scale to zero more cheaply. Here is the cost comparison.
Burstable T-family instances earn CPU credits when idle and spend them when busy, but under sustained load they exhaust credits and either throttle or, in unlimited mode, charge extra. Understanding when unlimited mode saves money and when it silently inflates the bill is worth knowing. Here is how it works.
ECS capacity providers decide where your tasks run and in what mix, and that decision drives the cost of an ECS cluster. Splitting between on-demand and Spot, tuning target capacity, and choosing Fargate versus EC2 providers can cut container spend substantially. Here is how they work on cost.
AWS Batch schedules containerized batch jobs across compute you configure, and because most batch work is interruption-tolerant, running it on Spot can cut compute cost by up to 90 percent. Here is how to configure Batch for Spot savings without losing jobs to interruptions.
S3 Glacier Instant Retrieval stores data at about $0.004 per GB, roughly a sixth of S3 Standard, while still giving millisecond access. Here is what it costs, where the retrieval fees bite, and when it beats Standard-IA and the slower Glacier tiers.
S3 replication looks like a checkbox but bills on four separate lines: destination storage, PUT requests, inter-region data transfer, and optional Replication Time Control. Here is how each is priced and how to keep replication from doubling your storage bill.
S3 versioning protects against overwrites and deletes, but every version is a full object billed at the full storage rate. Without lifecycle rules to expire noncurrent versions, a versioned bucket can store several times the data you think it does. Here is how to keep it in check.
S3 multipart upload is efficient for large files, but abandoned or failed uploads leave parts stored and billing forever, invisible in the normal object listing. Here is how the request and storage costs work, and the one lifecycle rule that stops the silent leak.
EBS io2 Block Express delivers up to 256,000 IOPS and 4,000 MB/s per volume at the same price as regular io2. Here is how the tiered IOPS pricing works, what a high-performance database volume actually costs, and when gp3 is the smarter buy.
On gp3 you provision IOPS and throughput separately, and paying for the wrong one is a common waste. Here is how the two are priced, how to tell which your workload needs, and why a database and a log archive want opposite settings.
RDS storage autoscaling prevents out-of-space outages by growing your volume automatically, and the feature itself is free. The catch is that it never shrinks, so a one-time spike can leave you paying for oversized storage for the life of the instance. Here is how to use it without the ratchet.
RDS gp3 gives a free IOPS and throughput baseline for a flat per-GB price, while io1 charges separately for every provisioned IOPS. For most databases gp3 is dramatically cheaper. Here is the pricing, the break-even, and when io1 still earns its cost.
RDS Reserved Instances cut the hourly rate of a database by up to about 40 percent for a one-year term and 60 percent for three years, in exchange for a commitment. Here is how the payment options compare, where the size-flexibility applies, and how to avoid over-committing.
Aurora Global Database gives sub-second cross-region replication and fast disaster recovery, but it bills replicated write I/O, storage in every region, secondary instances, and cross-region transfer. Here is the full cost breakdown and how to keep a global database from multiplying your bill.
A DynamoDB Global Secondary Index has its own throughput and storage, and every write to the base table that touches indexed attributes is also billed against the GSI. Here is how GSI cost works, why projections matter, and how a few indexes can double your write bill.
DynamoDB offers on-demand backups, point-in-time recovery, and S3 exports, each priced differently. Here is what each costs per GB, when continuous backup is worth double the on-demand rate, and how restores and exports are billed.
DynamoDB Accelerator (DAX) is a managed in-memory cache that can cut read latency to microseconds and read cost on hot data, but it runs on always-on nodes with a three-node minimum for production. Here is what DAX costs and the read volume where it starts to pay for itself.
Redshift DC2 nodes bundle compute and fixed local storage, while RA3 nodes separate them and bill managed storage per GB. For growing datasets RA3 is usually cheaper; for small, compute-heavy clusters DC2 can still win. Here is the pricing and the crossover point.
Redshift concurrency scaling adds transient clusters to absorb query spikes, and it comes with a daily free credit that covers many workloads for nothing. Here is how the free tier works, what overflow costs, and how to keep concurrency scaling from becoming an uncapped bill.