Operational Insights

Some of the most useful things your logs can tell you aren't about attacks. A server stopped sending, a scheduled task has been failing quietly for a week, or volume halved after Tuesday's deploy. Operational Insights writes those up as a short briefing, in sentences, with the evidence underneath.

Free 14-Day Trial
**No credit card required.

A briefing, not a dashboard

Open the page and the first thing you read is a list of findings, each in a sentence and each with a severity: Critical, High, Watch or Clear. Every finding has a button that takes you to the detail behind it: the logs, the sources involved, or the Security review.

Nothing is padded to fill the page. When the period was quiet, it says so:

Every source is reporting, and nothing failed.

That sentence is worth more than an empty chart, because it tells you the checks ran and found nothing, not that something is broken.

What it looks for

Finding What it means
Signs an attack may have got through A login that succeeded from an address straight after a brute force attempt from that same address, or an event log being cleared. Always Critical, always at the top, and linked to the Security Insights checks that cover them.
Sources that stopped sending A host or log source that was reporting has gone quiet: six hours or more on an hourly view, two days or more on a daily one. Only sources with meaningful volume count, so a host that logs twice a day won't cry wolf.
Availability trouble Hosts running out of memory or disk, services stopping, unexpected reboots.
Scheduled task and service failures Tasks and services failing to start or authenticate, usually because a stored password changed. These fail quietly and can go unnoticed for months.
Volume shift Log volume moved 25% or more between the first and second half of the period. Drops rank higher than rises, because something going quiet is usually the more important change.
New sources Systems that started reporting during the period, whether you added them or not.
Errors How many error events were recorded, with a link to the detail.
Attack attempts How many attack attempts were recorded, and what happened to them. Attempts stopped by an edge service or a mail filter in front of you are reported as blocked. Attempts that appear only in your own logs are reported as seen: we recognise them by pattern, but we see the request, not the outcome.

Silence is the finding that matters most

When a server stops sending logs, nothing happens. No error, no alert, no red light. The logs just aren't there, and you find out the day you need them.

That's why sources that stopped sending sit at the top of the briefing, ranked by how much they normally send. Only signs that an attack may have got through rank above them. If several stop at the same time, the page says it's probably one network path or configuration change, not several separate failures. That points you at the firewall rule or the syslog relay, not each server in turn.

It's also a compliance requirement. PCI DSS 10.7.2 requires that failures of critical security controls, including audit logging itself, are detected and acted on. Operational Insights shows it; the Logging Stopped alert tells you by email or Slack when an active host goes quiet. Our PCI DSS guide covers the rest of Requirement 10.

Volume that moved

A 25% change in log volume is rarely random. A drop can mean a log path changed in a deploy, an agent died, or a service stopped doing work. A rise can mean a new error loop, a debug flag left on, or traffic you didn't expect.

Operational Insights compares the median of the second half of the period with the first. Using the median means one noisy hour doesn't count as a trend. It tells you which way volume moved and by how much.

The evidence underneath

Below the briefing:

  • Events over time, per hour or per day, with the median marked so you can see what normal looks like.
  • Busiest sources, with each one's share of the total and a marker on any that have gone silent.
  • Noisiest addresses and countries, for when a change in volume is really a change in who's talking to you.

Works on any log source

Security Insights needs authentication logs and Web Insights needs access logs. Operational Insights needs neither. It works on whatever you send: Linux and Windows servers, firewalls, routers, applications. So it's useful on day one, even for an account sending nothing but syslog from network gear.

Common questions

How is this different from an alert?

An alert tells you about one event as it happens. Operational Insights tells you what changed over a period: today, yesterday, or the last 3, 7, 30, 60 or 90 days. Use both: the Logging Stopped alert for the moment a host goes quiet, and this page for the weekly look at what drifted.

Will it flag a server I switched off on purpose?

Yes, once, and that's deliberate. If you retired it, the finding is your confirmation. If you didn't, it's the most important line on the page.

Do I need to tell it what normal looks like?

No. It compares the period with itself (first half against second half, each source against its own usual volume), so there are no baselines to set.

Is it included in my plan?

Yes, on every plan. See pricing.

Simple, Affordable, Log Management and Analysis.

14 days free trial. No credit card required.