FeaturesCoveragePricingSecurity
FR EN
Features

Everything you need to truly monitor

From simple checks to transactional journeys, with a verdict confirmed by quorum and diagnostics that dig deep.

Eight monitor types

One monitor = one check. Each one knows exactly what it verifies and what triggers an outage.

URL / HTTP
Checks the response code, an expected keyword in the page, latency, followed redirects and TLS certificate validity.
Down on unexpected code, timeout, missing keyword or TLS error.
Ping (ICMP)
Reachability and network latency (average RTT), with optional packet-loss and delay thresholds.
Down if the host doesn't respond (100% loss) or a threshold is crossed.
DNS / Domain
Name resolution and registration watch via RDAP: expiry date, days left, registrar.
Down if unregistered, expired, or expiring under your alert threshold.
TLS certificate
Chain of trust, hostname match and expiry date; issuer and days left shown.
Down if untrusted, expired or expiring under the threshold — to alert you ahead of time.
TCP port
Opens a TCP connection on a port (SSH, SMTP, database, business port) and measures connect latency.
Down if the port is closed, filtered or times out; network diagnostics attached.
Transactional journey
Multi-step flow in a real Chromium (navigate, fill, click, assert_text…), recorded in a click with the recorder.
Down at the first failing step; screenshots (screen, console, source) attached.
Multi-step API
Chains HTTP calls: login, token extraction (JSONPath/regex) reinjected into the next step, rich assertions (code, time, JSON, header).
Down if an assertion fails; secrets and tokens redacted everywhere.
Full Page Check
Loads a page in a real browser and measures Web Vitals (LCP, FCP, CLS, TTFB), load time, page weight and request count.
Down if a threshold (load, LCP) is exceeded — or measure without alerting if left blank.

A fair verdict, not a reflex

Kolosili strictly tells a real outage from a measurement error — so nobody is woken up for nothing.

Up (UP)
The target responds and every check passes.
Down (DOWN)
The target doesn't respond or a check fails — confirmed by adaptive spatial quorum min(3, n) before any alert.
Measurement error
The monitor itself couldn't conclude (invalid config, broken JSONPath…). Not an outage: no alert is raised.
Unknown
Not enough measurements yet to decide (new or resumed monitor).

Three modes, from draft to production

Every monitor carries a mode that decides whether it measures and alerts — and what is billed.

01
Development
No automatic checks (only « Check now »), notifications off. Free — ideal to fine-tune a configuration.
02
Simulation
Regular checks but silent alerts. To validate behavior before going live.
03
Production
Fully operational: regular checks, active notifications, counted in SLAs.

Beyond « the site responds »

Multi-checkpoint & quorum
Every check runs from several checkpoints. An outage is confirmed by adaptive spatial quorum — min(3, n) checkpoints must see it — never on a single opinion. A single checkpoint's local error is not taken for an outage.
Multi-level alerting
1→3 escalation with promotion delays, reminders (interval + quota), per-level message and traceroute, email / SMS / webhook channels. Acknowledge stops the escalation; recovery notice on return.
SLA reporting & exports
Reproducible uptime reports (uptime, p50/p90/p95/p99, incidents), calendar or rolling periods, PDF / Excel exports and scheduled email delivery. Only Production counts in SLAs.
Incident diagnostics
Timed DNS / TCP / TLS phases, traceroute, resource waterfall and screenshots attached to every incident — to tell a real outage from a probe error.

In detail

Dashboards & observability
Per-user configurable widget dashboard, included in the plan: status, response time, SLA gauges, MTTR/MTBF, heatmaps, Top N, Web Vitals, per-step breakdown.
Global filters (period, group, monitors, checkpoint) and automatic refresh.
Save a layout as a report, export it to PDF/Excel or schedule its delivery.
Integrated Grafana observability for your Prometheus metrics (/metrics), isolated per company.
Recorder & maintenance
Scenario recorder: record a journey in a click in a driven browser, paste your secrets safely.
« Test scenario »: live run on a checkpoint, steps ✅/❌ as they go, without leaving a trace in history.
Scheduled maintenance (fixed or weekly recurring): alerts suspended and period excluded from SLA.
Security & governance
Secrets encrypted with AES-256-GCM, never returned or logged, redacted in diagnostic artifacts.
Strict multi-tenant isolation; fine-grained RBAC by roles, user groups and control ACLs.
Multi-tenant enterprise portals driven by a super-user; audit log of sensitive actions.
Day to day
Conversational AI assistant: natural-language questions, answers grounded in your own data, guided navigation — without ever changing anything for you.
Check details: steps, diagnostic artifacts, headers, network waterfall.
Clone a monitor (config, checkpoints, scenario) without ever exposing secrets.

Ready to stop chasing false positives?

See Kolosili on your own services, with your own checkpoints. A 20-minute demo is enough.