AWS-only managed key-value and document database with fixed per-partition throughput limits and no ad hoc queries.

etcd
A distributed key-value store using Raft consensus, built for cluster coordination rather than application data.
As of 13 November 2026, etcd is free to use. etcd gives you strongly consistent, linearizable reads and writes across a small cluster, and is the store behind every Kubernetes cluster. Softwr lists it under Databases.
Overview
What etcd does
etcd is a distributed key-value store written in Go, created at CoreOS, licensed under Apache 2.0 and now a graduated project of the Cloud Native Computing Foundation. It uses the Raft consensus algorithm to replicate a consistent log across an odd-numbered cluster, normally three or five members, and serves a gRPC API with linearizable reads by default. Its primitives are the ones coordination needs: atomic compare-and-swap transactions, leases with time-to-live, and watches that stream key changes to clients from a given revision. It is the datastore behind Kubernetes, and also underpins products such as CoreDNS setups, Patroni-managed Postgres and various schedulers. The distinguishing property is that it prioritises correctness over throughput and is honest about it. Every write is committed to a quorum and fsynced to disk before it is acknowledged, so a client that receives a success knows the write survives the loss of a minority of nodes, and a linearizable read never returns stale data. That guarantee is what makes it safe to build leader election, distributed locks and cluster membership on top of, which is the entire reason it exists; systems that need those primitives cannot use an eventually consistent store without inventing their own consensus. It is chosen by platform teams building schedulers, control planes and clustered systems, and inherited by everyone running Kubernetes whether they chose it or not. The trade-offs are the direct price of the guarantee. Performance is bounded by the slowest disk in the cluster because of the fsync on every write; adding members increases availability but decreases write throughput; and the store is sized for coordination data, with a default backend quota of 2 GB and 8 GB the recommended maximum, beyond which the cluster enters a read-only alarm state that a human must clear.
What people use it for
- Storing Kubernetes cluster state, which is what the overwhelming majority of etcd deployments are doing
- Leader election and distributed locking in a home-grown scheduler or control plane, using leases and transactions
- Service discovery and dynamic configuration where readers need to be notified of changes rather than poll for them
- Coordinating failover in a clustered database, as Patroni does for PostgreSQL
The honest half
Where it falls short
Concrete and checkable, so you can decide whether any of them matter to you. This is the half of a review a vendor will not write about etcd.
- It is sized for coordination data, not application data: the default backend quota is 2 GB and 8 GB is the documented recommended maximum, and exceeding it puts the cluster into a NOSPACE alarm where it accepts no writes until an operator compacts, defragments and clears the alarm by hand.
- Every write is replicated and fsynced before acknowledgement, so cluster performance is bounded by the slowest disk in it; a member on network-attached storage with high fsync latency causes leader elections and cluster-wide latency spikes that look like network problems and are not.
- Adding members increases availability but reduces write throughput, because each write must reach a larger quorum; you run three or five members for fault tolerance, and increasing capacity means faster hardware rather than more nodes.
- There is no sharding and no multi-tenancy, so isolating workloads means running separate clusters, each with its own quorum, certificates, backup schedule and upgrade path, and that operational multiplication is often unexpected.
- Running it yourself is a real job: periodic compaction and defragmentation, snapshot backups you have actually rehearsed restoring, and rotation of both peer and client TLS certificates, none of which happens automatically outside a managed Kubernetes service.
- Losing quorum is not self-healing; recovering a cluster that has lost a majority means restoring from a snapshot and accepting that everything written since that snapshot is gone, which makes backup frequency a data-loss budget decision rather than a routine setting.
Capabilities
Features
Raft consensus
Replicates a consistent log across members with automatic leader election, tolerating loss of a minority of nodes
Linearizable reads
Reads go through the leader by default and never return stale data, with a serialisable option when speed matters more
Transactions
Compare-and-swap style if/then/else transactions over multiple keys, which is what makes distributed locking safe
Leases
Keys with a time-to-live that a client must renew, so a crashed client's registration or lock expires automatically
Watches
Streams key changes to clients from a given revision, so a watcher that reconnects does not miss events
MVCC revision history
Keeps historical revisions so clients can read a consistent past snapshot, subject to the compaction policy
Role-based access control
Users, roles and key-range permissions, with mutual TLS for both client and peer connections
Snapshot backup and restore
A point-in-time snapshot command and a documented restore path, which is the recovery mechanism after quorum loss
Answered, with sources
Questions people ask
Each answer names the page it came from, so you can check it rather than take our word for it.
Can I use etcd as an application database?
No. It is designed for metadata and coordination, with a recommended maximum store size of around 8 GB, no sharding and a write path deliberately optimised for durability rather than throughput. Application data belongs in a database built for it.
How many members should a cluster have?
Three for most cases, five where you need to survive two simultaneous failures. Always an odd number, because an even-sized cluster gains no additional fault tolerance while making quorum harder to reach.
What happens when the store fills up?
The cluster raises a NOSPACE alarm and stops accepting writes, becoming effectively read-only. Recovery requires compacting old revisions, defragmenting each member and then explicitly disarming the alarm, all done by an operator.
Why is my etcd cluster slow or unstable?
Almost always disk latency. Because every write is fsynced before acknowledgement, slow or shared storage causes heartbeat timeouts, leader elections and cascading latency. Local SSDs with low fsync latency are effectively a requirement.
How does it compare with Consul or ZooKeeper?
All three provide consensus-backed coordination. etcd has the simplest data model and the Kubernetes ecosystem behind it; Consul bundles service discovery, health checking and a service mesh; ZooKeeper is older, JVM-based and still common under Kafka and Hadoop-era systems.
Checked against
Sources
The pages the facts above were read from. Prices and limits change, so the vendor's own page is the final word.
- etcd.ioetcd.io
- raw.githubusercontent.comraw.githubusercontent.com
Keep looking
Where to go from etcd
- etcd alternativesEvery product people weigh against it
- etcd pricingPlans, floors and what the tiers exclude
- etcd badgesWhich rules it meets, and the figure behind each
- All Databases softwareThe full category
- Best Databases softwareThe shortlist, and how it was cut
Best Databases software for
Compare etcd with
- DynamoDB
- Timeplus
- Firebase Realtime Database
- Knack
- Turso
- BigQuery
- TiDB
- Dgraph
- turbopuffer
- Apache Kafka
- Redis
- Postgres
- Presto
- Privacera
- RavenDB
- Readyset
- Redpanda
- Apache Airflow
Other Databases software
- Free, then $0.625/million writes10 researched notes
Streaming SQL engine built on ClickHouse internals, shipping as one small binary
Free, then $199/mo12 researched notesStore and sync data in real-time across all clients
Free plan8 researched notesHosted no-code database and web app builder priced by stored records and storage, with unlimited user accounts on every plan.
From $59/mo11 researched notesSQLite-lineage database platform for running very large numbers of small per-tenant databases, with an MIT-licensed engine.
Free, then $4.99/mo10 researched notesGoogle Cloud's serverless analytical warehouse, billed either by bytes scanned per query or by reserved compute slots.
Free plan13 researched notesApache 2.0 distributed SQL database with MySQL wire compatibility and a separate columnar replica for analytical queries.
Free plan12 researched notesApache 2.0 distributed graph database written in Go, maintained by Hypermode, queried through GraphQL or its own DQL language.
Free plan10 researched notesClosed-source vector and full-text search service built directly on object storage, with cold queries measured in seconds rather than milliseconds.
From $16/mo10 researched notesThe Meta-lineage distributed SQL query engine, distinct from the Trino fork
Open source12 researched notesCentralised data access governance from the creators of Apache Ranger, now rebranding as Trust3 AI
Pricing on request13 researched notesDatabase caching and optimization that reduces infrastructure costs 30-70%
Free plan10 researched notesProgrammatically author, schedule and monitor data workflows in Python
Open source11 researched notes
Softwr does not host reviews and shows no star rating for etcd, because a rating we did not collect is not ours to publish. What is here is the pricing and platform detail from the vendor’s own pages, limitations we could state concretely, and alternatives a reviewer confirmed people weigh against it. Tell us if any of it is wrong.
More on etcd
etcd alternatives
What people switch to, and what they give up
etcd pricing
Every tier, and where the cost actually lands
Compare etcd
Put it head to head with anything we hold
etcd badge
Its rating, and an embed for your own site
Best Databases software alternatives
Industrial process historian with published per-tag pricing and no client licence fees
Per tag band per yearNetApp-owned managed service for Cassandra, Kafka, OpenSearch, PostgreSQL and Cadence with a bring-your-own-cloud model
quoteErlang MQTT broker for large IoT fleets, relicensed to BSL with production free use limited to one node
Per month by connection and session volumeAttribute-based access control and masking applied inside Snowflake, Databricks and BigQuery
quoteMPP analytical database with a MySQL wire protocol and sub-second aggregation on wide tables
Open source, no licence feeStreaming database that maintains incremental materialised views in SQL instead of Flink jobs
Per RisingWave Unit hourCentralised data access governance from the creators of Apache Ranger, now rebranding as Trust3 AI
quoteThe Meta-lineage distributed SQL query engine, distinct from the Trino fork
Open source, no licence fee























