Skip to content

Logs & Log Search

GridNMS can collect the log messages your devices and systems produce, keep them searchable in one place, and watch them for patterns worth alerting on — like repeated login failures or a service crash. This page explains how to send logs in, how to search them, and how to turn log patterns into events.

The Logs page The Logs page lets you search across collected logs and filter by device, time, and text.

Devices send their logs to GridNMS using syslog, the standard logging protocol that almost every network device, firewall, and server already supports.

To start collecting logs from a device:

  1. On the device, point its syslog output at the IP address of the collector that serves its network, using the standard syslog port.
  2. Send a test message (or just wait for normal activity).
  3. The messages begin appearing on the Logs page in GridNMS.

For the full setup — ports, what a collector accepts, and verifying messages are arriving — see Receiving Syslog & Traps.

Open Investigate → Logs to search everything that’s been collected. The page is built around a single search bar, with a few controls above it:

  • Time Range — pick a window with the time-range picker (last hour, last day, custom). Searching a narrow window is much faster than a wide one.
  • Index — choose Operational logs (the default), Include system (_internal), or System only (_internal). GridNMS’s own housekeeping messages are kept out of the default view so your devices’ logs aren’t buried.
  • Saved — re-run a search you saved earlier, or choose Save current search… to keep the one you’re looking at. Saved searches are private to you.
  • Add to dashboard — put the current search on a dashboard as a panel.

In the search bar, type words to match, filters such as host=fw01, or pipe into a command such as | stats count() by host; suggestions appear as you type. A plain word finds that word anywhere in the message and is fast on large time ranges. Wrap a search in “double quotes” to match an exact phrase or part of a word. Press Enter or Run to search, and the ? button beside Run opens a short syntax reminder. Your recent searches are available from the Search history button.

Results come back newest first, each line badged with its own severity — from Informational up to Fatal. This is a different scale from the one on the Events page, where Critical is the lowest number, so read the name rather than the number when you move between the two: see Severity & Status Codes. The Fields panel summarizes the values across everything that matched — click a value to add it to your search — and Query Info shows how the search was run. A device’s own Logs tab on its detail page shows the same search pre-filtered to that device.

Combine the pieces to answer questions like “show me everything from edge-fw-2 containing denied in the last six hours.”

A device’s Logs tab is the same search, pre-filtered to that one device. The filter is wider than it looks, because the name a machine writes into its own log messages and the name you recorded on the device are two independent things and routinely disagree. A line is shown on the tab when any of these identifies the device:

  • the log’s host name matches one of the device’s names exactly;
  • the log arrived from one of the device’s addresses, or was relayed from one;
  • the log is already attributed to that device;
  • the host name matches one of its names ignoring case;
  • the first part of the host name, before the first dot, matches.

So a device recorded as web-node-01.corp.example still shows logs from a host writing WEB-NODE-01. The last rule is the loosest — a different host that happens to share the first part of the name can appear on the tab — so when you need certainty about which machine a line came from, read the host field on the line itself.

If a device you know is logging shows nothing, search the Logs page for its host name on its own. Finding the lines there but not on the device’s tab means the name it logs under is not recorded on the device; add it as the device’s DNS name and the tab finds it.

Searching is reactive — you only find a problem when you go looking. To get told automatically, set up a detection: a rule that watches incoming logs and raises an event the moment a message matches. GridNMS ships with built-in detections and you can add your own — see Detections.

Logs are easier to search and alert on when their values are named fields — like a source IP or username — instead of raw text. GridNMS extracts common fields automatically, and you can add your own; see Field Extraction.

Collected logs are stored for a retention period and then aged out automatically. Your retention is part of your plan, and you can see your current storage usage against your quota. If you’re collecting a lot of log volume, keep an eye on storage so you understand what you’re using.

Step What happens
Collect Devices send syslog to a collector; messages land in GridNMS.
Search You find messages on the Logs page or a device’s Logs tab.
Detect Detections watch incoming logs and raise events on a match.
Notify Those events flow to people via Notifications.

docs built 2026-11-12 · 945a2dec2