The product
Every server and application, one view.
Syncio watches what you run, explains what is slow or broken in plain language, and fixes what you allow it to fix — with approvals, receipts and undo.
Project health
86
-4 today
- Disk filling on web-02+5
- TLS certificate expires in 9 days+3
- Backup job skipped twice this week+2
- Everything else is in range0
- One install
- A single command on any Linux server, x86 or ARM, with no dependencies to add.
- One view
- Every server, project and organization on one screen, arranged the way your team is.
- One sentence
- Ask in plain English and get the cause, not a chart to interpret.
- One record
- Every change leaves a receipt, a verification step and a way back.
Four parts, one job.
Each part does one thing well, and hands the next one a clean answer.
A runtime that knows your server.
Install it in about a minute on any Linux server. It maps what is running, learns what normal looks like, and reports back.
- Measures what actually causes outages: CPU, memory, disk, network, services, certificates and backups.
- Runs fully on its own without an account, in local mode.
- Read-only by default; you choose how much access to grant, per server.
- Nothing leaves the server until you connect it and approve what is sent.
Install it in about a minute on any Linux server. It maps what is running, learns what normal looks like, and reports back.
Organizations, projects and servers.
One place to see and control every server and project, arranged the way teams and clients are.
- An organization for a team or a client; projects inside it for each system.
- Members, roles and invitations, per organization and per project.
- Health, jams and misses across everything, with drill-down to evidence.
- Settings inherit down the hierarchy, and every change is recorded.
4 servers
- web-01nginxHealthy
- web-02nginxDisk 91%
- db-01postgresHealthy
- worker-01jobsBacklog
The score, and the reasons behind it.
A 0 to 100 score for every server, project and organization, with the top reasons it is not 100 and how many points each fix would recover.
- Health: availability, security, recoverability, capacity, maintenance and performance.
- Jams: anything slow, stuck or saturated — CPU waiting in line, slow disks, choked networks.
- Misses: anything that failed or did not happen — failed requests, missed backups, expired certificates.
- Every tile opens the evidence behind it.
Project health
86
-4 today
- Disk filling on web-02+5
- TLS certificate expires in 9 days+3
- Backup job skipped twice this week+2
- Everything else is in range0
Ask. Approve. Done.Beta
Ask about your infrastructure in plain English and get an answer with the evidence behind it. Tell it what to do and it prepares the action for your approval.
- It knows the page you are on, the project and the server, and your role.
- It uses the same permissions you have — never more.
- Anything that changes something waits for your approval and leaves a receipt.
- It can explain a chart, an incident, a forecast or a score change.
YouWhy is checkout slow since this morning?
A calm briefing, and reports clients read.Beta
A daily or weekly summary of what happened and what needs you, by email or chat, plus branded reports you can send on a schedule.
- The weekly Pre-mortem ranks the few things most likely to break next.
- Reports carry health, incidents, security and performance.
- Branded and scheduled, for clients and managers.
Ranked by risk
- Disk on web-02 will fill in about 6 daysfix it
- Certificate renews in 9 days, no automationfix it
- Backups verified twice this weekfine
Show your customers what is going on.Beta
Public or private status pages, mapped to your real checks, with drafted updates you approve before they publish.
- Components map to checks and servers; status follows real data.
- Subscribers by email, SMS, RSS or webhook, with double opt-in.
- A custom domain and your own branding.
- Websiteoperational
- Image deliverydegraded
- APIoperational
What it measures, and why each one matters.
Not a dump of every metric the kernel exposes. These are the things that actually cause a bad night, each with the reason it is on the list.
Availability
- Uptime and reachability
- Whether the server and its services answer, checked continuously rather than on a schedule you have to set.
- Certificate expiry
- A cert that lapses takes the site down. It is counted as a miss the moment it is close, not the hour it dies.
- Backup freshness
- A backup that stopped running is invisible until you need it. Its last success is a first-class number here.
Capacity
- CPU, and CPU waiting in line
- The load average and the time processes spend queued tells you a server is saturated before users notice.
- Memory and swap
- Swap in use is the early sign of a leak or a limit set wrong, long before the out-of-memory killer arrives.
- Disk, space and speed
- A full disk and a slow disk are different failures, so both are measured and reported separately.
Performance
- Requests and latency
- Slow endpoints and error rates, tied to the server conditions behind them rather than floating on their own.
- Network in, out and saturation
- Bandwidth used, packets dropped and connections refused — the shape of a network problem, not just its size.
- Services and processes
- Which units are running, which restarted, and which are eating the machine right now.
Change
- Deploys and config changes
- What moved, and when — so the answer to "what broke it" is on the same timeline as the symptom.
- Topology
- What talks to what, learned from the traffic, so an outage can be traced across servers and not just within one.
- Security posture
- Stale packages, open ports, exposed services and leaked secrets, confirmed from the outside where possible.
What happens when you approve a fix.
Nothing here is a black box. A change moves through the same six steps every time, and you can stop it at any of them.
- 01
Ask or propose
You tell the Copilot what to do, or it proposes a fix from a finding. Either way it writes out the exact command and the reason.
- 02
Read it first
The change is shown in full before anything runs: what it touches, what it expects, and what it will do if the server disagrees.
- 03
Approve
A person approves it. The Copilot can propose, but it cannot approve its own work — approval is always a human action.
- 04
Snapshot
Before the first write, the relevant state is captured, so there is something concrete to return to.
- 05
Run and verify
The change runs, then the same checks that found the problem are run again to confirm it actually worked.
- 06
Receipt and undo
The whole thing is recorded as a receipt with the before, the after and a one-click way back.
Pending your approval
Clear disk on web-02 by rotating logs older than 14 days.
Read-only by default. Nothing runs until a person approves it, and every approval leaves a receipt with an undo.
The score, and the reasons behind it.
Health is a 0 to 100 number with the points each problem is costing you. Here is what a real project view looks like — the score first, then the reasons, each one opening onto its own evidence.
Project health
86
-4 today
- Disk filling on web-02+5
- TLS certificate expires in 9 days+3
- Backup job skipped twice this week+2
- Everything else is in range0
Ask. Approve. Done.
The Copilot knows the page you are on, the project and the server, and your role — so a question can be short and the answer can be specific. It uses the same permissions you have, never more.
- Explain a chart, a spike or a score change in a sentence.
- Find what changed right before a problem started.
- Draft the fix and show it to you before it runs.
- Write the incident note while the timeline is fresh.
- Answer in the same terms you already use for your stack.
It sits beside what you already run.
Syncio is not trying to replace your metrics, your logs or your on-call. It reads across them, explains the cause, and adds the one thing they tend to leave out: a safe way to act.
- Metrics and dashboards
- Keep them. Syncio answers the question the chart raises, and points back at the data behind the answer.
- Log aggregation
- Keep it. Patterns from logs become findings here, with the server conditions that produced them attached.
- Uptime checks
- Keep them. Syncio measures reachability too, and ties a failure to the change that caused it.
- Incident and paging tools
- Keep them. Alerts route where you already work; the timeline and the receipt come from here.
- Configuration management
- Keep it. Syncio does not own your desired state; it watches what is actually true and flags the drift.
- Your own tooling
- The API and MCP let your scripts and assistants read the same view a person sees in the dashboard.
You decide how much it can do.
Set an access level per server, and lower it at any time. There is also a switch on the server itself.
- 01ObserveRead-only. Syncio can see, not change.
- 02OperateApproved fixes, each with a receipt. Beta
- 03ConsoleAn interactive terminal for you, recorded. Beta
- 04AgentThe Copilot may run approved commands for you. Early Access
Get started
See it against your stack.
Syncio explains causes and can fix with approval. Keep your charts and add answers.