What Epok doesn't do
Epok picks one job and tries to do it well: turning your signals — logs, metrics, traces, RUM, and session replay — into a verified incident with a drafted root-cause hypothesis. Everything else, we route you to a better tool.
This list is the honest scope statement. Most observability vendors imply they do everything; we'd rather tell you where we end and where someone else begins.
Host-level infrastructure RCA from logs alone (CPU / disk / memory / network at the node level)
Why we don't: These faults rarely emit useful application logs. The kernel, hypervisor, and cloud control plane log to themselves, not into your app's log stream. You can push host metrics to Epok over Prometheus remote_write or OTLP — and pull cloud-provider metrics (CloudWatch, Cloud Monitoring) in via the OpenTelemetry collector's receivers — and we'll watch them; but we are not a replacement for a full infrastructure-metrics stack with node-level dashboards.
Use instead: Prometheus + node_exporter, or your cloud provider's native metrics (CloudWatch, Stackdriver, Azure Monitor).
Synthetic monitoring / uptime checks (HTTP pings, TLS cert checks)
Why we don't: We watch your logs after a request happens. We don't probe your endpoints to find out when they're down. We can detect a service going silent in its logs but that's after the failure, not before — though for failures that build gradually (saturation, capacity, exhaustion) Epok does forecast the breach ahead of time from your metrics; it just doesn't actively probe endpoints.
Use instead: Better Stack, Pingdom, UptimeRobot, Checkly.
SDK-instrumented in-app exception capture
Why we don't: When your Rails / Django / Express app throws an exception, Sentry's SDK captures it before the response is rendered, with full stack frames and breadcrumbs. We see whatever your logger emitted. If your code doesn't log the exception, we don't see it.
Use instead: Sentry, Rollbar, or your language-native exception tracker.
Compliance + SIEM (SOC 2 audit logs, GDPR investigation tooling, security incident response)
Why we don't: We are not a SIEM. We don't have the retention guarantees, immutable audit trail, role-based access boundaries, or compliance certifications that a security-incident pipeline requires.
Use instead: Splunk, Datadog Security, Panther, or your cloud provider's native security center.
What we do, briefly
Anomaly detection across every detector layer spanning logs, metrics, and traces, with the incident correlated across every signal you send — logs, metrics, traces, RUM, and session replay — on one canvas by shared trace ID, with cross-service causal confirmation and one incident summary per outage instead of 47 raw alerts. Ingest is standard push protocols — no proprietary agent to install. See the home page for the long version.