Cuando una web va lenta, la pregunta es siempre la misma: ¿es la máquina, que se ha quedado pequeña, o es otra cosa? AWS tiene la respuesta en CloudWatch, pero en forma de gráficas con nombres como CPUCreditBalance y NetworkOut que hay que saber interpretar. Y la métrica que más veces explica el problema en las máquinas pequeñas, la capacidad de ráfaga, es justo la que casi nadie mira.
Simplio lee esas métricas de la nube, sin instalar nada en la máquina, y las resume en frases: "esta máquina va justa", "queda poca capacidad de ráfaga", "todo en orden". Debajo, las gráficas de las últimas 24 horas para quien quiera el detalle. Lo mismo para las bases de datos: CPU, conexiones y espacio libre.
How it works
You open the machine's card
The metrics are on the card itself: CPU, burst capacity (on machines that have it) and inbound and outbound network traffic, for the last 24 hours.
You read the sentence
Simplio interprets the values and tells you in one sentence how the machine is doing. If it's running tight, it tells you before the website goes down.
If you want, you look at the chart
The charts are underneath, to compare hours and see whether the problem is a one-off or constant.
How secure it is
Nothing installed on your machine
The metrics come from AWS (CloudWatch), with read-only permission. No agent, no program on the server.
The same task, in AWS and in Simplio
| On your own in AWS | In Simplio | |
|---|---|---|
| Know whether the machine is running tight | CloudWatch → metrics → choose instance, period and statistic → interpret CPUCreditBalance | The sentence on the card |
| See yesterday's network | Same route, another metric | The 24 h chart under the sentence |
| Databases | Another service, other charts | Same card, same sentences |
Frequently asked questions
What is burst capacity?
Small AWS machines (the T family on EC2 and most of Lightsail) can't use 100 % of the CPU all the time: they accumulate "credits" while idle and spend them under load. If they run out, the machine becomes slow even if the CPU chart doesn't look like it. The guide "My machine is slow" explains it from scratch.
Does it alert me if something goes wrong?
The metrics are checked on the card. Email alerts today cover other things (stopped machine, missing copy, expiring domain…).