Monitoring integrations come with a predefined set of charts and dashboards. Great if you want to get started quickly and don’t really know which metrics you should be monitoring or don’t want to spend time building charts and dashboards from scratch. Now, in addition to these out of the box charts and dashboards you can also craft completely custom charts for monitoring your applications using all available metrics. These charts can be added to dashboards to visualize your infrastructure and application behavior.Custom charts give you full control over how you want to visualize your data. Some data series are easier to analyze when presented as bar charts, others are better rendered as line charts. You can also customize colors, line styles, stacking, and more. It’s all to make your monitoring easier. You decide which presentation method gives you the most power. Ship metrics to Sematext with input format compatible with Influx and build your own dashboards with any metrics by its name. Imagine you want to correlate CPU and memory usage. The easiest way to see the correlation is to draw CPU and memory usage on the same chart/ CPU usage is expressed in percentages while memory is measured in bytes. No worries – just configure two Y-axes units and assign CPU usage to one and memory usage to the other Y-axis, as shown in this short video:
Do you want to group data series by hostname and sum them up to get total memory usage in your cluster? Just pick the tag to group by and the aggregation function. Do you want to see memory usage of each host separately? That’s just as easy: pick a grouping tag and no aggregation function (‘all separately’).
Custom charts are a beta feature for now. Over the coming weeks you’ll be able to:
- Perform math operations on your data. For example, you’ll be able to create a data series derived from two or more metrics.
- Use filter variables, so all charts on the dashboard can be controlled interactively with a single click without the need to change each chart configuration separately.
- Use the query editor if you prefer typing over clicking to create your charts.
- Building charts that combine logs counts and metrics on a single chart.
Core infrastructure monitoring with Sematext just got better. How? The new agent collects container metrics and events just like the old agent, but it also collects Kubernetes metrics and events. As a result, you can now see a number of Kubernetes dashboards out of the box.You can now identify processes that use the most resources, whether CPU or memory, across your entire infrastructure. Forget about ssh-ing to machines and running top. In Sematext you can see the hungriest processes right now, but also their behavior over time. Processes running in containers are also monitored, giving you visibility inside containers as well. The new inventory monitoring captures information about installed packages, their versions, operating system information, Linux distributions, and many other useful bits. Package install and uninstall events details are also captured, so you can always trace who changes what, where, and when. Using Sematext you now have a place where you can identify outdated packages, packages with vulnerabilities, groups of servers with identical sets of packages as well as those machines that you thought have the same setup but in fact don’t, and more. Network monitoring was rewritten to utilize eBPF whenever possible instead of relying on network packets. This new implementation has almost no overhead. The agent was rewritten in Go, shedding the need for the JVM, resulting in smaller CPU and memory footprint. All of the above-mentioned reports are a part of the new type of Monitoring App we are calling Infra App. You can have any number of Infra Apps in your Sematext account, though most people will have one Infra App per monitored environment. For example, you might create one Infra App for monitoring all of your production applications and infrastructure and another for monitoring your staging environment. Learn more about infrastructure monitoring.
If installing the Sematext Agent on Kubernetes manually felt too old fashioned and if Sematext Agent Helm Chart was not good enough, you’ll be pleased to hear that Sematext Agent is now available as a Kubernetes Operator. Sematext Operator is available on Github at https://github.com/sematext/sematext-operator. You can also find it certified in the RedHat’s OpenShift Container Platform via OperatorHub.io. Learn more in Kubernetes agent docs.
Do you know how many different Linux distros you are running in production? Do you know which ones you are running and their versions? How can you tell which version of openssh is installed on your servers? Or maybe you are running several different versions and you can’t really easily check that? Do you know if you are running any software with known vulnerabilities?With Sematext’s new Inventory Monitoring you can now answer such questions with ease. The new infrastructure Inventory gives you information about your operating systems, Linux distributions, hardware specs, installed packages, log files, and more. The agent also detects package installation and removal events. Your event stream shows which packages were installed, removed, or updated, by who, where, and when. All inventory data is collected by the Sematext Agent and is part of core infrastructure monitoring. To start seeing your inventory information in Sematext be sure you are running a recent version of Sematext Agent. For more info check out inventory monitoring documentation and the official announcement.
You can now use Sematext for real-time process monitoring. With process monitoring you gain real-time visibility into which processes use the most CPU or memory, whether they are running inside containers, in cloud instances, your own VMs, or bare-metal servers. To start collecting information about processes make sure you are using a recent Sematext Agent. For more information see process monitoring documentation and the official announcement.
You can now control during which times of the day and which days of the week alert rules are active or inactive. By default an alert rule is active all day, every day. The new alert scheduling lets you pick, for each day of the week, one or more time slots when the alert rule is active. Each alert rule can have its own schedule. Each day of the week can have different times when an alert rule is active. You can have different schedule Monday to Friday vs. Weekend, for example. The alert schedule can be modified at any time and it comes into effect immediately. Have a look!