10 DynamoDB Advantages and 10 Disadvantages
DynamoDB scales to millions of requests with no DBA, but the 400KB item cap and rigid access patterns bite. Ten upsides, ten tradeoffs, and when to walk.
16 articles in this category
DynamoDB scales to millions of requests with no DBA, but the 400KB item cap and rigid access patterns bite. Ten upsides, ten tradeoffs, and when to walk.
The scalar, document, set and binary types DynamoDB supports, how each counts against the 400KB item limit, and the empty-value rules that break writes.
Items cap at 400KB, transactions are ACID, and there is no open-source build. Short answers on limits, consistency, indexes, scaling and DynamoDB billing.
A GSI gives your table a second partition key so you stop scanning. Projections, eventual consistency, capacity planning, and sparse and inverted indexes.
Thirty questions interviewers actually ask, answered: partition keys, single-table design, capacity units, adaptive capacity, Streams, DAX and GSIs.
DAX is a write-through cache that drops DynamoDB reads from milliseconds to microseconds. How the item and query caches work, what it costs, when to skip.
The partition key decides where an item lives, the sort key decides its order within. How to choose both, and the hot-partition mistakes to avoid.
PITR restores a table to any second in the last 35 days. How to switch it on, how a restore really runs, what it costs, and where on-demand backups win.
Streams records a before and after image of every write, kept for 24 hours. How shards and Lambda triggers work, plus replication and aggregation patterns.
DynamoDB is AWS's serverless key-value store: single-digit millisecond reads at any scale. How partitions and keys work, what it costs, when to avoid it.
Ten rules that decide whether DynamoDB runs fast and cheap or slow and expensive: access-pattern design, key cardinality, index restraint, TTL.
An LSI shares the partition key and must exist from table creation. A GSI can be added later, at its own cost. Which to reach for, and what to project.
Work out what a table will actually cost. On-demand versus provisioned compared, an interactive calculator, and where the bill usually leaks.
PutItem replaces the entire item. UpdateItem changes only the attributes you name. When each is correct, and the silent data loss PutItem can cause.
Working Query examples for the AWS CLI, Boto3 and the Node SDK: sort key conditions, GSI queries, pagination, projections and fixes for empty result sets.
25GB of storage plus 25 read and 25 write capacity units, free every month. The exact limits, the charges that catch people out, and how to stay inside.