24/7 monitoring agent that ships fixes for disk, memory, CPU, and Kubernetes incidents — and escalates only what it can't handle.
Last shipped: 2026-05-30 — v0.1.30Pick a real incident type. BytePort detects it, picks a runbook, remediates, verifies, and writes a postmortem — in under 30 seconds. Demo environment — no real infrastructure is touched.
Every card below is a live runbook — a real action BytePort takes when a signal breaches threshold. Not a recommendation. A fix.
Browse the full runbook catalog →Vacuums journalctl logs, rotates stale log files, prunes Docker layers, sweeps /tmp. Resolves without restart in 90% of cases.
Identifies the leaking process via cgroup stats. Restarts it using the byteport.restart=safe label so only opted-in services are touched.
Applies cgroup CPU throttle via byteport.throttle=safe label. Buys headroom without killing the process, then files a postmortem with root cause.
Runs connectivity checks, service probes, and NTP drift detection. Restarts unhealthy services in dependency order, escalates if the host itself is unreachable.
Classifies crash cause (OOM, config error, missing dep). Safe-restarts the pod with back-off guard. Requires byteport.io/allow-remediation: "true" annotation on the namespace.
Reads the byteport.io/allow-rollback annotation. If set, triggers kubectl rollout undo and notifies the commit author via Slack or email.
Distinguishes memory leak from under-provisioned limits. If limits look correct, restarts with debug logging enabled. If leak signature detected, patches the resource limit and files a postmortem.
Point it at your signal source. Give it a notification target. It watches from there.
npm install @byteport/agent
# minimum required env vars
export BYTEPORT_API_KEY=your_api_key
export PROMETHEUS_URL=http://prometheus:9090 # or DATADOG_API_KEY
export POSTMARK_SERVER_TOKEN=your_token # or SLACK_WEBHOOK_URL
node -e "require('@byteport/agent').start()"docker run -d \
-e BYTEPORT_API_KEY=your_api_key \
-e PROMETHEUS_URL=http://prometheus:9090 \
-e POSTMARK_SERVER_TOKEN=your_token \
byteport/agent:latesthelm repo add byteport https://byteport.polsia.app/charts
helm repo update
# Default install is dry-run safe — never mutates without your permission
helm install byteport byteport/byteport-agent \
--namespace byteport-agent \
--create-namespace
# Confirm it's watching (logs appear within 30s)
kubectl logs -n byteport-agent \
-l app.kubernetes.io/name=byteport-agent --tail=20 -fPlug into your existing signal sources and notification channels. No new dashboards required.
Real incidents from the live database. Updated as they happen.
BytePort is in private beta. Join the waitlist and we'll reach out when your stack is a good fit.