Palo Alto Networks SD-WAN monitoring for every site and hub
Firewall Fleet Health monitors Palo Alto Networks SD-WAN by reading latency, jitter and loss for every path, the state of every tunnel, IKE failures and BGP peers from each firewall. It lays them out as a hub-by-site matrix and a tunnel drops heatmap, so one bad link or one misbehaving hub stands out. It keeps 7 days of 5-minute detail and 13 months of daily trends.
What does it measure on each SD-WAN path?
Every sweep reads the path monitor on each firewall: latency, jitter and loss for each path, over the three averaging windows the firewall keeps (high, medium and low sensitivity). Probe loss is worked out from the probe requests and replies between sweeps, so a path that quietly drops probes shows up even while its averages look calm.
That is per-path SD-WAN latency, jitter and loss history, the kind of view usually sold as a premium cloud add-on, included here and kept as daily trends for 13 months.
How do I see every site against every hub?
The hub matrix puts every site in a row and every hub in a column, with paths up out of paths total in each cell. HA pairs are merged into one site. A cell turns yellow when a path is down or the tunnels between that site and hub drop more than 1.5 times an hour.
| Site | Chicago-DC | Dallas-DC | Azure-Hub |
|---|---|---|---|
| Akron | 2/2 | 2/2 | 1/1 |
| Boise | 2/2 | 2/2 | 1/1 |
| Columbus | 2/2 | 1/2 (a path is down) | 1/1 |
| Dayton | 2/2 | 2/2 | 1/1 |
| Eugene | 2/2 | 2/2 | 1/1 |
Is the hub or the site causing the tunnel drops?
The tunnel drops heatmap shows drops for each site over time, charged to the far-end site, and hovering a cell lists the firewalls that logged them. A hub is blamed only when it is the cause, for example when three or more sites have trouble with the same hub, or the hub's own paths are all down. When one site drops its tunnels to two different hubs within two minutes, the problem is that site's own, so a single flaky branch does not paint its hub yellow.
Tunnel status is kept for 30 days, enough to see whether this week's drops started last week.
What about IKE failures and BGP?
IKE failures come from the firewalls' VPN logs and are charged to the right firewall. When a hub logs a stale child SA that does not name the tunnel, and exactly one branch's tunnels went up or down on that hub within 15 seconds, the failure is charged to that branch.
BGP peers are read on every sweep, with their state and flap count, so a peer that bounces between sweeps is still counted. Tunnel drops and IKE failures alert at 20 in 24 hours, scaled to the range you are looking at.
What else does it show for each site?
Alongside the SD-WAN views, each firewall's Device page shows the rest of the picture an engineer needs.
- IPsec tunnels and tunnel monitors
- WAN throughput, interface errors and drops
- Bandwidth by application for the site's links
- HA state and sync, and the last reboot with its reason
See it on a fleet
Open the SD-WAN tab on the demo fleet, or book a walkthrough and see it on your own sites.