A monitoring system is only as good as its alert channel. The prettiest dashboard in the world is useless at 2 a.m. when nobody has a laptop open. Alerts have to find the person on call, inside an app that person is guaranteed to open. In Vietnam, that app is Zalo.
What is Zalo, and why does it matter for alerts?
If you are reading this from outside Vietnam: Zalo is the country’s dominant messaging app, the local equivalent of WhatsApp in Europe or KakaoTalk in Korea. It is the default channel for family, coworkers and customer conversations, installed on virtually every phone and opened many times a day.
That usage pattern is exactly what an alert channel needs. Notifications for Zalo stay on with sound in places where email notifications were muted years ago. So if your ops team is in Vietnam, or works with partners there, routing incidents into Zalo means they land where people are already looking.
Why should alerts go to Zalo instead of email?
Email remains the standard channel for international monitoring tools, and it has exactly one fatal flaw: nobody reads email at midnight. An alert email queues up next to invoices, newsletters and social notifications, waiting for morning.
A Zalo message behaves differently. It appears on the lock screen, makes a sound, and most importantly goes to the shared team group: whoever sees it first takes it, and no incident sits unread in the inbox of someone on vacation.
This is also the sharpest difference between AgentWatch and international tools like UptimeRobot: they have no reason to build a Zalo channel, while for us it is the default. The UptimeRobot comparison digs further into that difference in focus.
How does AgentWatch deliver Zalo alerts?
AgentWatch uses Zalo Bot to post messages into a group chat you choose. The life of an incident looks like this:
- The system detects a problem: a server stops reporting, an HTTP check fails, a service goes down.
- An incident opens and an alert fires into the Zalo group, naming the server, the service and the start time.
- When the metric returns to normal, the incident closes itself and the group gets a recovery message with the total duration.
The recovery message matters as much as the alert. It tells the on-call person they can go back to sleep, and it gives the whole team one honest timeline per incident instead of three different memories of it.
How long does setup take?
A few minutes, done exactly once:
- Create an AgentWatch account if you do not have one.
- Create a Zalo Bot and add it to your ops team group.
- In the alert channel settings, paste the bot token and the group chat id, then register the webhook.
- Send a test message to confirm alerts reach the right group.
From then on, every incident from every server and check flows through this channel. Adding a new server requires no extra configuration.
Which incidents show up in Zalo?
Everything AgentWatch monitors can end as a Zalo message:
- Server unresponsive: the agent goes silent, the machine is off or the network dropped.
- Resources over threshold: CPU, RAM or disk crossing the limits you set.
- A systemd service stopped: a watched process is no longer active.
- A Docker container unhealthy: through the three-layer health model described in Docker monitoring.
- HTTP, TCP or ping check failing: a website or API no longer responds correctly.
- SSL certificate expiring soon: warned early enough to renew calmly.
What else besides instant alerts?
Per-incident alerts are half the story. A few extra mechanisms turn your Zalo group into a genuine ops hub:
- Daily digest: sent at times you choose, by default 07:00, 12:00 and 18:00, leading with anything unusual. On a quiet day it is one short ✅ message confirming the monitoring itself is alive.
- Ask the bot in the group: type a report command and get an on-the-spot summary without opening the dashboard.
- Unresolved incident reminders: anything open past 30 minutes gets re-sent, so nothing scrolls away.
- Maintenance windows: schedule quiet periods for planned work while history keeps recording.
What other alert channels exist?
Email already runs alongside Zalo, useful as a paper trail or for teammates who do not use Zalo. Telegram is in development and will launch during the beta. There is no Slack or Microsoft Teams integration, and we would rather say that plainly than promise it vaguely.
What does Zalo alerting cost?
Every AgentWatch plan, including Free with 2 servers and 10 service checks, gets the full Zalo and email channels. Paid plans from 49,000 VND per month differ in server count, check interval and history retention. Details are on the pricing table and the product overview on the homepage.
During the beta, new accounts run the Business plan free. It takes a few minutes to make the next alert land where your whole team is already looking: sign up and connect Zalo now.
