True observabilityat any scale.
Logs, metrics and traces in a single unified engine built on object storage:
fast, 10x cheaper, with semantic and free-text search.
Built by veterans from:
Freedom to go Berserk
Choosing an observability engine shouldn't mean replacing your incident tools, buying another AI subscription or moving data out of your cloud. Berserk works with your own agents and runs on your infrastructure, or ours. It can also run alongside the platforms you already have, for the data that costs too much to keep in them. You decide where the data lives and how your team investigates it.
Keep approved AI
Use the models and agents you've already approved and paid for. We don't resell AI tokens.
Choose where data lives
Run Berserk on-premises, in your private cloud, or let us host it. Compatible with most cloud providers.
Keep your process
Query and investigate through your terminal, the UI, your agent or any custom workflow.
Cost calculator
Run it on-prem or in your private cloud.
Ingest per month
1 TB
Hardware requirements
?6 cores
across the cluster
10 GB
RAM across the cluster
100 GB
object storage
23 GB
local cache (NVMe)
Sized by our hardware sizing guide
Hardware on AWS (estimate)*
40¢ / GB
~€400
Berserk License
FREE
* Hardware estimate based on Amazon AWS list prices: c6id instances with local NVMe plus S3 for retention, on demand. The cache has to fit on the instances' NVMe, so the estimate can buy more vCPU than the cores above. An example, not a quote.
Scales to any size
Built on object storage
Berserk scales from your first terabyte to petabytes a day, across every region you run in. Compute and storage scale independently, and a highly optimized ingest path keeps up with whatever you send.
1.0 TB/month
