Explore our new Cybersecurity Terms Glossary, explained by our experts.
Cybersecurity Terms Glossary
Coverage is the product of two things: which event types you collect, and what proportion of machines you collect them from. Most organisations overestimate both, and only discover it during an incident.

What Is Endpoint Telemetry Coverage?
Endpoint telemetry coverage answers two questions that organisations usually conflate. Which classes of event do you collect, and from how much of the estate do you collect them? A tool deployed to 94% of machines that records process creation but not script execution has a coverage problem in both dimensions, and the gaps are invisible until someone goes looking for evidence that was never captured.
| Event class | What it supports |
|---|---|
| Process creation with command line and parent | Most behavioural detection |
| Network connections attributed to a process | Command-and-control, exfiltration |
| File creation and modification | Ransomware, staging, persistence |
| Registry modification | Persistence, defence evasion |
| Module and driver loading | Injection, rootkits, unsigned code |
| Script block content | Obfuscated PowerShell and scripting |
| Authentication and privilege use | Lateral movement, credential abuse |
| DNS queries with process attribution | Resolution of malicious infrastructure |
Process creation with the full command line is the foundation, which is why command-line auditing is the first thing to confirm. Without parent process information, a large part of behavioural detection becomes guesswork, since the signal is frequently the relationship rather than the process itself.
Agent coverage is a simple calculation that organisations consistently get wrong, because the denominator is taken from the same inventory the agent populates. Compare at least three independent sources: the directory, the network access control or DHCP records, and the agent console. The machines appearing in the first two and not the third are the ones that matter.
Then measure liveness rather than installation. An agent installed but not reporting for 11 days is worse than no agent, because the console shows green. Alert on any endpoint that has sent no telemetry within an expected interval, and treat sustained silence as an incident on high-value systems.
Finally, measure event coverage per platform. The common pattern is strong Windows workstation coverage, weaker Windows server coverage because change control slowed the rollout, thin Linux coverage, almost nothing on macOS, and nothing at all on the appliances. Record the percentage by platform and event class, and attach it to your detection coverage map, because a rule covering 55% of hosts is not the same as a rule.
Telemetry generated on a host is not telemetry available to you. The chain of agent, forwarder, network path, ingestion pipeline, parser, and index has several failure points, and each one fails quietly. Measure the volume per host per day and alert on deviation, verify that parsing still maps fields correctly after any platform upgrade, and use automated detection testing to prove the whole path end to end rather than inspecting the parts.
Coverage costs money, in licences and in ingestion volume, so the sequence matters. Confirm process and command-line telemetry everywhere before adding exotic event types. Prioritise platforms by what an attacker would target rather than by what is easiest to instrument, which usually promotes servers and domain controllers above laptops. And collect broadly but tier storage, keeping the full detail hot for a short window and the rest in an archive, a decision covered under retention planning.
At Privia Security, we begin detection work by measuring actual coverage against three independent inventories, because the figure in the console is almost never the figure on the ground. Our SOC maturation services include that reconciliation.
Aim for 98% or better of managed endpoints reporting within the last 24 hours, measured against an inventory the agent does not build. The last few per cent are usually a mix of decommissioned machines still in the directory, devices that never return to the network, and systems deliberately excluded. Each exclusion should have an owner and a compensating control rather than being an unexplained gap.
Process creation including the full command line and the parent process. It supports the largest number of detections, from living-off-the-land abuse to ransomware precursors, and without the parent relationship much behavioural detection cannot be written at all. Script block logging for PowerShell is the usual second priority on Windows, since it defeats most command-line obfuscation.
Use the sources those systems do produce. Network telemetry gives you connection records and, with a sensor, protocol detail. The devices themselves usually emit syslog, which is worth collecting even when it is sparse. Configuration monitoring detects unauthorised change. For operational technology, passive network monitoring purpose-built for industrial protocols is normally the only acceptable option.
Monitor expected volume per source and per host, and alert when a reporting endpoint goes quiet for longer than its normal gap. Add synthetic checks that deliberately generate a known event on a schedule and verify it arrives in the index. Relying on the agent console’s health view is insufficient, because an agent can report healthy while its events are dropped further along the pipeline.
You May Be Interested In These
We use a few cookies that are needed to make this website work, and they are always on. We would also like to set analytics cookies, to see which pages are useful; functional cookies, so that our live chat works; and advertising cookies, to measure our campaigns. We will not set any of those until you say yes, and you can change your mind at any time from the Cookie Settings link in the footer.
Choose the cookies you are happy for us to use. Strictly necessary cookies are always on; everything else is off until you turn it on.
These cookies keep the website working: they hold your security checks, your form submissions and the choice you make here. The site cannot run without them, so they are always on.
These let our live chat open and remember your conversation, and let the newsletter sign-up work. Switch them off and the chat window will not appear.
These tell us which pages are read and how people find us, so that we can improve the site. The reports are about groups of visitors, never about you personally.
These let us measure our advertising and show you our adverts on other websites, such as LinkedIn, Google and Meta. Switch them off and you will still see adverts, but they will be less relevant.