Built for the deployments that don't make the public roadmap.
Air-gapped clusters. BYO-cloud with your KMS keys. Signed package repos. Multi-region failover. A named SRE on call. Tell us what your environment looks like — we'll tell you what we already ship and what we'll build.
Every question already has an answer.
The questions a senior platform engineer asks before a single pod ships. We answered them when we built the agent, not after a sales call.
Does our data touch your servers?
No. Compression happens on the agent inside your cluster; compressed chunks go directly from the agent to your bucket. We handle routing config and a training model — metadata only.
Who holds the keys?
You do. Your KMS keys, your bucket policy, your VPC endpoint. Our agent never gets long-lived credentials; it authenticates via your cloud’s native identity federation.
Can we revoke access instantly?
Yes. Named API keys per environment (prod, staging, CI). Revoking a key stops that deployment in under a second without touching the others. No approval workflow, no support ticket.
What happens when we double our cluster?
DaemonSet — one agent per node. Cluster grows, agents grow with it. No central broker, no bottleneck, no per-agent license.
How do we monitor the agent itself?
/health endpoint per pod, Prometheus metrics, structured JSON logs. Integrates with whatever you already use to watch DaemonSets. The dashboard shows live ratio + upload lag per agent.
How do we roll back?
Delete the Helm release. Your log pipeline is exactly where it was — we deployed alongside, never in front. No data rewriting, no queue drain, no coordination meeting.
The #1 unblocker for platform-team approval. Answered before they ask.
Do we have to uninstall the Datadog / Splunk / New Relic / Dynatrace agent we already run?
No. That agent keeps running for APM, metrics, NPM, code-level tracing, security — everything it already does. You flip log collection off on that agent with a single config flag, then add Sasquatch as a separate DaemonSet that handles only logs. Two agents on the same node, no conflict, reversible in seconds. Nobody at your company has to approve ripping out your observability stack.
logs_enabled: falseremove filelog receiverlogs: enabled: falsedisable Log Monitoring per host grouphelm uninstall.