Ping
Check that a host answers ICMP echo requests from every region you pick, and watch its round-trip time and packet loss.
How the ping check works
A Ping monitor sends ICMP echo requests to a hostname or IP address and waits for the replies. Each region you selected sends its own packets on every check, so a host that answers from Europe but not from Asia shows up as a regional failure. The region results are combined as described in Core Concepts: the monitor goes down when more than half of its regions fail.
In each region:
- The check sends Packet count echo requests (4 by default) and waits up to Timeout (10 seconds by default) for each reply.
- It passes when at least one reply comes back. Lost packets don't fail the check; they are reported as packet loss.
- It fails when no reply comes back at all, including when the hostname doesn't resolve.
- The response time recorded is the average round-trip time of the replies. The minimum, maximum, jitter (the standard deviation of the round-trip times) and packet loss are kept with each check.
Create a Ping monitor
- 1Choose the type
On the Monitors page, click New, open the type picker and choose Ping monitor. It is listed under Website & availability.
- 2Enter the host
Fill in Hostname or IP address with a hostname or an IPv4 or IPv6 address, such as
api.example.comor203.0.113.10. No scheme, no port and no path. - 3Adjust the ICMP settings if you need to
Open Advanced settings. The ICMP settings section sets how many packets are sent and how long to wait for them (see the table below). IP version, in the same panel, chooses between IPv4 and IPv6.
- 4Create it
Pick who is alerted under How should we escalate incidents? and click Create monitor. If you leave the name empty, the monitor is named after the host.
Settings
| Setting | Options | Default | What it does |
|---|---|---|---|
| Packet count | 1, 2, 4 or 8 packets | 4 packets | Echo requests sent per check, in each region. More packets make the packet-loss figure more meaningful and make the check slower. |
| Timeout | 5, 10, 20 or 30 seconds | 10 seconds | How long to wait for each reply. A reply that arrives later counts as lost. |
| TTL (Time to Live) | 32, 64, 128 or 255 hops | 64 hops | The packets' time to live. A host more hops away than this never receives them, and the check fails. |
| Packet size | 32, 56, 128 or 512 bytes | 56 bytes | The ICMP payload in each packet. |
| IP version | IPv4 + IPv6, IPv4 only, IPv6 only | IPv4 + IPv6 | Which address family to ping. IPv4 + IPv6 leaves the choice to the system; the other two force one family, so a host without an address of that family fails. |
Through the API, packetCount accepts 1 to 10, ttl 1 to 255 and packetSize 16 to 1,472 bytes; values outside those ranges are clamped. Check frequency, Regions and the confirmation and recovery periods work as for every monitor; see Monitor Types.
When it goes down
| What happened | Message on the check |
|---|---|
| No reply to any packet before the timeout — the host is down, ICMP is blocked, the name doesn't resolve, or the host is beyond the TTL | Request timed out |
| The host is, or resolves to, a private, loopback or other internal address | db.example.com resolves to 10.0.0.5, which is a private network address and cannot be checked from Zharp's servers |
A passing check reads Ping successful: 23ms, with the average round-trip time. Partial packet loss doesn't fail a check: a check where one of four packets is answered passes with 75% loss. Watch the packet loss column and the Packet loss (24h) figure on the monitor page for a host that is reachable but unreliable.
Internal addresses are refused when you save the monitor, and again at every check in case the name has since been pointed at one. To watch a host that is only reachable inside your network, see Private Networks.
The monitor page
A Ping monitor's page shows:
- A bar with the host, the packets sent per check and the TTL of the last reply, marked REACHABLE or UNREACHABLE (PENDING before the first check). The TTL shown is what was left on the reply when it arrived, not the setting, so it hints at how far away the host is.
- A Round-trip time chart.
- Recent checks, with Status, Host, RTT, Packet loss, Location (the region) and Date. Packet loss above zero is shown in red.
- Expanding a check shows the Resolved IP, the packets sent and received, the packet loss, the jitter, and the MIN, AVG and MAX round-trip times.
- The Ping settings card, with the Resolved IP, Packets / check, the Timeout from the ICMP settings, the check Interval and Packet loss (24h): the average loss over the last day, in amber at 5% or more.
Troubleshooting
The host is up but the check says Request timed out
The most common cause is a firewall that drops ICMP. Many cloud providers block inbound ping by default — AWS security groups, Azure network security groups and Google Cloud firewall rules all need an explicit rule — and some hosts and CDNs ignore echo requests on purpose. Allow inbound ICMP echo requests (type 8) and outbound echo replies (type 0) for the host, or monitor a TCP port on it with a Port monitor instead.
Also check that:
- Hostname or IP address holds only the host. A URL such as
https://example.com/is not something ping can reach. - IP version matches the addresses the host has. IPv6 only on a host without an IPv6 address always fails.
- TTL (Time to Live) is high enough. With 32 hops, a host on the far side of the world can be out of reach; 64 is enough for almost every route.
Only one region fails
The host is reachable, but not along that region's route — a regional network problem, or a firewall or geo-block that only admits some sources. While the other regions pass, the monitor stays up — its page shows the failing region, and nobody is alerted. Expand a failing row to see which region it came from.
Packet loss is high but the monitor stays up
A check passes as long as one packet is answered. Some routers rate-limit ICMP, which shows up as steady loss on an otherwise healthy host. Raise Packet count to get a steadier figure, and compare the loss across regions: loss from every region points at the host or its network, loss from one region at the route.
Large packets fail, small ones pass
Something on the path drops large or fragmented ICMP packets. Keep Packet size at the default 56 bytes unless you are deliberately testing the path's MTU.