Skip to content
Tutorials Center
My learning

Topics

Lesson 24 of 37 ยท Dashboards and Reporting

Graphs, Maps and Screens

Learn the three kinds of Zabbix graphs, how to build network maps with link indicators, how screens became dashboards, and how Nagios graphs via add-ons.

Numbers in a table tell you the current value; graphs tell you the story. Maps show how things connect and which part is on fire. This lesson covers the graph types in Zabbix, how to build maps with live status, and what happened to "screens", a term you will still find in older guides. It ends with how Nagios gets graphs. The product cannot run on this site, so the stand-ins compute the same ideas in Python.

Three kinds of graphs

  1. Simple graphs: free, one per numeric item, available from Latest data. No configuration. Useful for a quick look.
  2. Custom graphs: configured on a host or template; combine several items, choose type (normal, stacked, pie, exploded), colors, Y-axis limits, and draw style.
  3. Ad-hoc graphs: select items in Latest data, choose "Display graph" or "Display stacked graph" for a quick comparison.

Custom graphs can also be graph prototypes, created per discovered object (see discovery rules).

Dashboards offer the Graph widget, which replaced most use of classic graphs. It supports data sets with patterns (vfs.fs.size[*,pused]), overrides per series, left and right Y axes, threshold lines for triggers, and time shift to compare "now" against "one week ago".

Reading a graph correctly

  • Trends versus history. For short ranges Zabbix draws raw history. For long ranges it draws trends: hourly min, avg and max per item. If you reduced trend storage in the item (for example to 7 days), a 30 day graph is empty.
  • Y-axis scale. Auto scaling hides how small a change is when the axis starts above zero. Set min and max where "percent used" should run 0 to 100.
  • Missing data. A gap means no value arrived, not zero. Use "Show working time" and "Show triggers" options deliberately.
  • Units. Bits per second versus bytes per second is a classic mistake: net.if.in returns bytes, so a multiplier of 8 is needed for bits.

Network maps

A map is a drawing with elements and links. Open Monitoring > Maps and click Create map. Elements can be:

  • Host: shows problem state through colored icon and label.
  • Host group: aggregates all its hosts.
  • Trigger: shows a single condition.
  • Image: static decoration such as a building.
  • Map: a link to a sub-map for drill-down.
  • Shape and line: pure drawing.

A link between two elements can have indicators: a link turns red when a chosen trigger fires. With a trigger on interface status or link utilization, the map shows broken links directly. Labels can use macros such as {HOST.NAME} or {?last(/host/net.if.in[eth0])} to display a live value.

Map building tips:

  1. Start from a real topology (WAN, data center, application tiers), not from the host list.
  2. Use sub-maps. A top map shows sites; click a site to see the rack or the service.
  3. Use "Automatic icon selection" to show different icons by host inventory type.
  4. Keep the number of elements per map low enough to read; 30 is a lot.
  5. Map navigation trees on a dashboard let people switch between maps.

The map Geomap widget is a separate thing: it places hosts on a world map from the inventory fields location_lat and location_lon.

What happened to screens

Older Zabbix versions (up to 4.x) had screens: grids that combined graphs, maps and text. Since 5.4 screens were merged into dashboards, and upgrades convert existing screens automatically. If you read a guide about "Configuration > Screens", translate it to "Dashboards" and "Graph" widgets. Slide shows became dashboards with several pages.

Nagios graphs

Nagios Core's own interface has a status map (the classic "Map" view) and trend reports, but no time-series storage. Performance data is added to plugin output after the | character:

TEXT
DISK OK - free space: / 14320 MB (65%) | /=7668MB;21500;24187;0;26875

Tools that read this performance data build the graphs: PNP4Nagios (RRD files), Graphite or InfluxDB with Grafana through a small broker module, or Nagios XI which bundles graphing. Check the format label=value[unit];warn;crit;min;max and you can parse it yourself.

Stand-in: parse Nagios performance data

Python Parse plugin performance data (Python stand-in)

Press Run (or Ctrl + Enter) to see the result here.

Program input ยท what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with input()), it gets the next line from this box.

Zabbix keeps hourly trends: min, avg and max of the raw values. This stand-in computes them so you can see why a short spike remains visible in the max column even after raw history is deleted.

Python History to hourly trends (Python stand-in)

Press Run (or Ctrl + Enter) to see the result here.

Program input ยท what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with input()), it gets the next line from this box.

The average of hour 1 hardly moves, while the max column keeps the spike.

Common mistakes

  • Reading a 90 day graph after setting trend storage to 7 days.
  • Building maps with raw host lists and no topology, which makes them unreadable.
  • Using bytes where bits were meant, then wondering why the link looks only 12.5 percent used.

Check your understanding

Quick check ยท 4 questions

of 4 answered You scored / 4 ยท perfect!

  1. 1. What do trends store?

    Trends are hourly aggregates used for long-range graphs.
  2. 2. What replaced screens in modern Zabbix?

    Screens were merged into dashboards, which also took over slide shows.
  3. 3. How does a map link show that a connection is broken?

    Link indicators change color when a chosen trigger is in problem state.
  4. 4. Where does Nagios plugin performance data appear?

    Format: label=value;warn;crit;min;max after the | in the output line.

Summary

  • Use simple, custom and ad-hoc graphs, and the dashboard Graph widget for modern views.
  • Long-range graphs use trends, so trend retention limits how far back you can look.
  • Maps combine elements and links with trigger-based indicators; build them from topology.
  • Screens are gone; dashboards replaced them. Nagios graphs rely on performance data and add-ons.

Next: SLA and Availability Reports