Loading the SOTA2 catalog…
ObsessionDB
Fully managed hosting for ClickHouse software and a data warehouse for AI, with compute-storage separation and deployment across AWS, Google Cloud, and Azure.
- Founded
- Not disclosed
- Headquarters
- Not disclosed
- Company size
- 1–10
- Products tracked
- 1
Products by ObsessionDB
Pricing for ObsessionDB
Free, Usage Based, Enterprise Custom
This is the broad commercial model reported for ObsessionDB. Product-level terms are listed below when available.
- Pricing model
- Free, Usage Based, Enterprise Custom
- Entry option
- Free plan and Free trial
- Starting price
- Free developer tier
Trust profile
Security and compliance
No compliance claims have been published for this company.
Tenant isolation with separate Kubernetes namespaces, compute, object-storage buckets, coordination databases, credentials, and default-deny network policies
TLS encryption in transit
Encryption at rest via storage-provider protections; server-side object-storage encryption rollout in progress
Role-based access controls with table- and column-level scoping
Individual query logging and auditability
Dedicated compute and private networking for enterprise workloads
Daily coordination-metadata backups with 14-day retention
Documented data deletion procedure targeting completion within 7 days
Incident notification within 72 hours of becoming aware of an affected personal-data breach
Latest updates from ObsessionDB
New feature
ObsessionDB adopts ClickHouse 26.8 LTS
ObsessionDB announced that it now runs ClickHouse 26.8 LTS and is fully compatible with the upstream release. The version is available to every customer and includes upstream capabilities such as faster aggregations and joins, background query execution through run_query_in_background, reduced Parquet I/O, and additional tokenizers for full-text search in languages including Chinese and Japanese.
New feature
ObsessionDB ships faster projection queries
ObsessionDB shipped engine and cache-layer improvements that make projection-backed point lookups faster for all customers. On a production table exceeding 200 billion rows with a 20+ TB projection, the company reports 213 ms p50 and 703 ms p99 latency across 1,000 consecutive runs, with no query exceeding one second. The changes keep projection metadata in dedicated in-memory caches and improve distributed-cache read handling.
Research
ObsessionDB reports sub-second projection queries at 200 billion rows
ObsessionDB published an engineering report on improving ClickHouse projection queries over object storage. On a six-node production cluster serving a table with more than 200 billion rows, the company kept projection metadata resident in memory and tuned the distributed cache path for small reads. It reports 213 ms median and 703 ms p99 latency across 1,000 runs, with the slowest query completing in 0.98 seconds.
MentionedClickHouse
Buyer request
Want to contact this company?
Tell SOTA2 how you would like to connect. We securely save your request for this company's verified team, whether the profile is claimed now or claimed later. Anonymous visitors stay anonymous.