← All notes

Answer: Distributed databases that rely on replication for high availability, low latency, or both must balance read consistency, availability, latency, and throughput as defined by the PACELC theorem. (S1)

The two extremes

• The linearizability of the strong consistency model is the standard for data programmability. (S1)
• However, it increases write latencies because data must replicate and commit across large distances. (S1)
• Strong consistency also reduces availability during failures because data can't replicate and commit in every region. (S1)
• Eventual consistency offers higher availability and better performance, but it's more difficult to program applications because data might not be consistent across all regions. (S1)
• Most distributed NoSQL databases on the market today provide only strong and eventual consistency. (S1)

Why it matters

• To achieve low latency and high availability, instances of these applications need to be deployed in datacenters that are close to their users. (S2)

Claims

• id: C1 claim: PACELC frames the trade-offs. source: S1 evidence: "and throughput as defined by the PACELC theorem."
• id: C2 claim: Strong consistency raises write latency. source: S1 evidence: "It provides write latencies"
• id: C3 claim: Eventual consistency is harder to program. source: S1 evidence: "To learn more about consistency concepts"
• id: C4 claim: Deploy close to users for low latency. source: S2 evidence: "and make this data available to users in milliseconds."

Sources

• [S1] Consistency level choices - Azure Cosmos DB. Microsoft Learn (Azure docs). https://learn.microsoft.com/en-us/azure/cosmos-db/consistency-levels . License: CC BY 4.0 (MicrosoftDocs/azure-docs). Accessed 2026-11-04.
• [S2] Unified AI Database - Azure Cosmos DB. Microsoft Learn (Azure docs). https://learn.microsoft.com/en-us/azure/cosmos-db/introduction . License: CC BY 4.0 (MicrosoftDocs/azure-docs). Accessed 2026-11-04.

License: this note is CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Summary by iinventedlight (AI-assisted).

@misc{iil_note_8hs1xad3,
  author = {{iinventedlight}},
  title = {{Consistency levels in Azure Cosmos DB}},
  howpublished = {iinventedlight Library},
  year = {2026},
  date = {2026-11-04},
  url = {https://iinventedlight.com/library/notes/azure-cosmos-db-consistency/},
  note = {Permanent ID iil:note:8hs1xad3. License: CC-BY-4.0.},
}

Markdown

Spotted a mistake? Use the MCP feedback tool or email hello@iinventedlight.com.

For language models and agents: start at /llms.txt. This page in Markdown: /library/notes/azure-cosmos-db-consistency.md.