anyChangelog
What shipped, when. This page is also the answer to "will this company still exist next year?"
Debian, CentOS Stream, openSUSE and Amazon Linux
2026-11-03
The same one-line install now works on Debian 12+, CentOS Stream 9+, openSUSE Leap 16+ and Amazon Linux 2023, on both x86-64 and arm64 — it picks the right package for the server. No curl on a fresh Debian? Use wget instead; the install page shows how.
anyglance is open
2026-10-31
Anyone can sign up. Use every feature on up to 5 servers for 30 days, then pick a plan or carry on with Free (2 servers).
Pay by card
2026-10-31
Pick a plan on the Account page and the Paddle checkout opens. Your card details never reach us. Cancelling, changing your payment method and your billing history are on the same page, and you can add servers one at a time.
Alerts on your phone
2026-10-27
Get alerts on iPhone and Android without installing an app. Turn on Phone in Settings and the same alerts as email show up on your lock screen; tap one to open that server. On iPhone, add anyglance to your Home Screen from Safari first. The server list, detail, alerts and settings pages are also tidied up for phone screens. Available on every plan.
Open alerts first
2026-10-25
Recent alerts now groups the ones still open at the top. Resolved ones sit below.
Size for every disk
2026-10-23
Each disk row on the detail page now shows used / total size (GB) next to the usage. Usage, size and fill-up forecast line up in columns, so you can see at a glance which disk fills up when.
Apache and IIS, before they fill up
2026-10-22
If Apache is on your watch list (Ubuntu and Debian), we show how many of its workers are busy; for IIS, how full the request queue is. No setup. You get an alert after 3 minutes above 80%, and right away with "Requests rejected" once IIS starts turning requests away.
Weekly report
2026-10-21
Every Monday morning, last week in one message: report rate per server, CPU and memory average and peak, disks with their fill-up forecast, alert counts, and agents on old versions. It goes to email and Slack, and you can turn it off in Settings.
nginx, before it fills up
2026-10-21
If nginx is on your watch list, we show what share of its connection limit is in use. No setup (Linux). You get an alert after 3 minutes above 80%.
Alerts in Slack too
2026-10-20
Add a Slack webhook in Settings and every alert email is also posted to Slack, with the same text. Send a test to check it right away. Available on every plan.
Collected every 10 seconds
2026-10-14
The collection interval went from 30 seconds to 10. Short CPU or connection spikes now stay on the chart. The price is unchanged.
Delete your account yourself
2026-10-12
From the Account page, with your password. Your account and every server, rule and token go immediately; collected metrics can no longer be viewed and are erased within 90 days. No support email needed.
Reboot or shut down from the browser
2026-10-12
Two fixed actions in Server settings, each confirmed by typing a word. The agent still runs no command you type — these two are all it accepts. The server isn't flagged as unreported for 10 minutes after a reboot.
Who is using the GPU
2026-10-11
Server detail shows which processes hold VRAM and how much, plus the driver version. When a Linux box lacks nvidia-smi, it tells you which package to install instead of showing empty cards.
Watch services on macOS, and ports everywhere
2026-10-11
On Mac the agent watches launchd services as it does systemd and Windows services. On every OS you can also name a port; the agent checks it on its own loopback, so nothing has to be opened to the internet.
GPU on Windows, any vendor
2026-10-10
Windows servers report GPU usage and memory through the OS itself, so NVIDIA, AMD and Intel cards all show up.
An OS section on server detail
2026-10-09
Load average, inodes, open files, and on Windows commit charge and processor queue — the numbers you check when the machine feels slow but CPU looks fine.
An alert when a watched service stops
2026-10-07
No rule to write. Pick the services to watch; if one stops for more than a minute you get an email, and the recovery email says which ones came back.
Clock offset, TCP connections, and threshold lines
2026-10-06
Two new metrics: how far the server's clock has drifted and how many TCP connections it holds. Charts now draw your alert thresholds as lines, so you see how close you are.
When a server doesn't connect, we say where it broke
2026-10-05
If no report arrives a few minutes after install, onboarding tells you whether the agent never reached us or reached us and stopped — with the command to check, per OS.
macOS support
2026-10-02
Apple Silicon Macs install with the same one-line command and run as a launchd daemon. Apple GPU usage is collected too.
NVIDIA GPU monitoring
2026-10-02
Per-card utilization, VRAM, temperature and power on Linux, next to CPU and disk on the same screen. Values the driver doesn't report stay blank instead of showing 0.
Disk projection
2026-10-01
Servers now show "full in ~N days" from the fill trend. 88% is a number; three days is a decision.
Bigger charts and per-core CPU
2026-09-29
Click any chart to open it large. CPU can be drawn per core, which shows the one pinned core an average hides.
One-line install on Linux and Windows
2026-09-25
A single command installs the agent with the token already in it — a checked .deb on Linux, an MSI on Windows that can also be pushed by GPO. If the Windows service dies, it restarts itself.
English and Korean
2026-09-23
Every screen comes in both languages. English is the default; your choice is remembered in the browser.