Intro To Longview
Linode Longview is a systems analytics platform that is provided FREE to all Linode customers. You install an “agent” on a server you would like to monitor CPU and memory utilization as well as Disk I/O and network interface performance. This agent is decently easy to install, most the time it is a single Copy/Paste from the Linode Cloud Manager into your Terminal.
Linode also has incredible documentation for how troubleshooting your Linode Longview agent. I’ve many-a-times needed to update the API key located in the following filepath…
/etc/linode/longview.key
What It Does Well
The ‘root’ view in the Cloud Manager nicely lays out 6 gauges for each server including hostname, uptime, last check-in time, and available package updates.
If you click into one of your servers you can easily see more details about the underlying system, laid out much like the Windows Task Manager you are likely familiar with.
What Longview Misses
I’ve played with a handful of monitoring platforms during my time in IT. SolarWinds, DataDog, UptimeKuma, NewRelic, Longview, and other open-source projects as well like Cockpit-pcp.
No Alerting – No really! You cannot setup automated emails or webhooks to be triggered based on predefined alerting conditions. This is a major fall-back for anyone serious about responsive incident management.
No ARM Support – While Raspberry Pi users are not the target audience, its still disappointing to see that ARM and RISC-V are not at all supported from a platform perspective. Developers and Home-Labbers alike might not be able to holistically monitor their Infrastructure using Longview.
No Logging Ingest – By not collecting and analyzing server logs inside your monitoring platform, you leave a service gap. I already mentioned that responsive incident management is hindered by no alerting but this also impacts proactive responses. Other platforms that compete in this space such as DataDog and NewRelic go that extra step and also collect System Logs from their agents. I find it disappointing that even paid plans cannot collect Logs from inside the platform.
My Summary
For my simple use-case of tracking spikes in server utilization and helping guesstimate when I need to perform OS patches, it does its job to my satisfaction. However due to its lacking features I do monitor my servers with other tooling. Paid plans do not exactly expand the feature set already available for Free.
Use it, or don’t… It’s okay for what it is.
Comments are closed