Live Logs

Live Logs Without
the Infrastructure

Stream, filter, and search logs from all your Kubernetes pods — no Loki, no Elasticsearch, no cloud account.

↓ Download for macOS Download for Windows

Live Logs — 3 services selected

INFWRNERR

14:32:01.234api-gatewayINFGET /api/v1/users 200 OK (12ms)

14:32:01.456auth-serviceINFToken validated for user_id=8291 scope=read,write

14:32:01.891api-gatewayINFGET /api/v1/projects 200 OK (34ms)

14:32:02.102payment-svcWRNStripe webhook retry attempt 2/5 for evt_1234

14:32:02.340api-gatewayINFPOST /api/v1/deploy 202 Accepted (8ms)

14:32:02.567auth-serviceINFSession refreshed for user_id=4102 ttl=3600s

14:32:03.001payment-svcERRStripe webhook failed: Connection timeout after 30s

14:32:03.210api-gatewayINFGET /health 200 OK (1ms)

14:32:03.445auth-serviceINFRate limit check: 42/1000 requests in current window

14:32:03.678payment-svcINFInvoice inv_5678 created amount=49.99 currency=EUR

14:32:04.012api-gatewayWRNSlow upstream response from payment-svc (2.1s)

14:32:04.234auth-serviceINFNew API key issued for org_id=301 name=ci-deploy

Streaming from 2 clusters, 4 namespaces

Buffer: 12,847 / 50,000 lines

Collection

stern-Based Collection

CPI-Control spawns background stern processes for each connected namespace. Logs are collected continuously into an in-memory ring buffer and streamed to the UI via Server-Sent Events. No log files on disk, no database writes, no ingestion pipeline.

✓Background stern processes — one per namespace per cluster
✓In-memory ring buffer with configurable size (default 50K lines)
✓SSE live streaming — logs appear in the UI within milliseconds
✓Automatic process recovery if stern crashes or disconnects

Log Collectors

production-eudefault (12 pods)

~340 lines/min

production-eumonitoring (4 pods)

~80 lines/min

staging-usdefault (8 pods)

~120 lines/min

staging-usbatch-jobs (3 pods)

~45 lines/min

Ring buffer: 12,847 / 50,000~585 lines/min total

Views

Multi-Service Views

Select multiple services and see their logs interleaved in real time. Each service gets a color-coded badge so you can instantly tell which service produced which line. Pause the stream, search through the buffer, or filter by log level — all without losing your place.

✓Select any combination of services across clusters
✓Color-coded service badges for instant identification
✓Pause/resume without losing incoming logs
✓Full-text search across the entire ring buffer

Selected Services

api-gatewayauth-servicepayment-svc

apiRequest processed in 12ms

authToken refresh completed

apiUpstream timeout: payment-svc

payRetry queue: 3 pending

authSession cache hit rate: 94%

payInvoice created: EUR 49.99

Window

Floating Log Window

Detach the log viewer into a separate window that stays on top. Monitor logs on a second monitor while you navigate dashboards, manage deployments, or investigate incidents in the main window. The floating window maintains its own scroll position and filter state.

✓One-click detach into a separate always-on-top window
✓Independent scroll position and filter state
✓Perfect for multi-monitor setups
✓Re-dock back into the main window at any time

CPI-Control — Dashboard

Service Dashboard

Floating Logs

api GET /health 200

auth Token validated

pay Invoice created

api Slow response 2.1s

Zero Overhead

No infrastructure to maintain

Traditional log aggregation means running Loki, Elasticsearch, or paying for Datadog's log ingestion. CPI-Control needs none of that.

No Disk Storage

Logs live in a ring buffer in memory. When the buffer is full, the oldest lines are evicted. No log rotation, no disk cleanup.

No Database

No Elasticsearch cluster. No ClickHouse. No TimescaleDB. The ring buffer is the only storage, and it resets when you restart.

No Cloud Account

No AWS CloudWatch, no GCP Cloud Logging, no Datadog log ingestion. Your logs never leave your machine.

No Maintenance

No index management, no retention policies, no storage capacity planning. Just connect your cluster and view logs.

Comparison

CPI-Control vs. log infrastructure

CPI-Control trades long-term log retention for zero operational overhead. If you need live debugging and tail-style log viewing, it does the job without the baggage.

	CPI-Control	Loki + Grafana	Datadog Logs
Setup time	0 minutes	Hours	Minutes
Infrastructure needed	None	Loki + Grafana + Promtail	Agent per host
Cost (50 services)	Free	Self-hosted cost	$200+/month
Multi-cluster	Built-in	Manual config	Built-in
Multi-service view	Yes	LogQL queries	Yes
Floating window	Yes	---	---
Data location	Your machine	Your cluster	US Cloud
Retention	Ring buffer (memory)	Configurable	15 days (default)

Logs from every pod.
Zero infrastructure.

Connect your clusters and start streaming logs immediately. No setup, no infrastructure, no monthly bills. Free for up to 50 services.